id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2409.10589 | Jesse Van Remmerden | Jesse van Remmerden, Zaharah Bukhsh and Yingqian Zhang | Offline Reinforcement Learning for Learning to Dispatch for Job Shop Scheduling | Accepted in Machine Learning | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The Job Shop Scheduling Problem (JSSP) is a complex combinatorial optimization problem. While online Reinforcement Learning (RL) has shown promise by quickly finding acceptable solutions for JSSP, it faces key limitations: it requires extensive training interactions from scratch leading to sample inefficiency, cannot leverage existing high-quality solutions from traditional methods like Constraint Programming (CP), and require simulated environments to train in, which are impracticable to build for complex scheduling environments. We introduce Offline Learned Dispatching (Offline-LD), an offline reinforcement learning approach for JSSP, which addresses these limitations by learning from historical scheduling data. Our approach is motivated by scenarios where historical scheduling data and expert solutions are available or scenarios where online training of RL approaches with simulated environments is impracticable. Offline-LD introduces maskable variants of two Q-learning methods, namely, Maskable Quantile Regression DQN (mQRDQN) and discrete maskable Soft Actor-Critic (d-mSAC), that are able to learn from historical data, through Conservative Q-Learning (CQL). Moreover, we present a novel entropy bonus modification for d-mSAC, for maskable action spaces. Moreover, we introduce a novel reward normalization method for JSSP in an offline RL setting. Our experiments demonstrate that Offline-LD outperforms online RL on both generated and benchmark instances when trained on only 100 solutions generated by CP. Notably, introducing noise to the expert dataset yields comparable or superior results to using the expert dataset, with the same amount of instances, a promising finding for real-world applications, where data is inherently noisy and imperfect. | [
{
"version": "v1",
"created": "Mon, 16 Sep 2024 15:18:10 GMT"
},
{
"version": "v2",
"created": "Wed, 08 Jan 2025 15:41:04 GMT"
},
{
"version": "v3",
"created": "Sun, 13 Apr 2025 14:52:43 GMT"
},
{
"version": "v4",
"created": "Thu, 03 Jul 2025 11:41:37 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"van Remmerden",
"Jesse",
""
],
[
"Bukhsh",
"Zaharah",
""
],
[
"Zhang",
"Yingqian",
""
]
]
|
2409.10672 | Tamal RoyChowdhury | Pratul Manna, Tamal RoyChowdhury, Chandra Kant Mishra | Improved inspiral-merger-ringdown model for BBHs on elliptical orbits | 18 pages, 12 figures | Phys. Rev. D 111, 124026 (2025) | 10.1103/849s-3zy8 | null | gr-qc astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gravitational waveforms capturing binary evolution through the early-inspiral phase play a critical role in extracting orbital features that nearly disappear during the late-inspiral and subsequent merger phase due to radiation reaction forces; for instance, the effect of orbital eccentricity. Phenomenological approaches that model compact binary mergers rely heavily on combining inputs from both analytical and numerical approaches to reduce the computational cost of generating templates for data analysis purposes. In a recent work, Chattaraj et al., Phys. Rev. D 106, 124008 (2022) constructed a dominant ($\ell=2$, $|m|=2$) mode model for nonspinning binary black holes (BBHs) on elliptical orbits. The model was constructed in time domain and is fully analytical. The current work is an attempt to improve this model by making a few important changes in our approach. The most significant of those involves identifying initial values of orbital parameters with which the inspiral part of the model is evolved. While the ingredients remain the same as in the previous work, the resulting (new) model, when compared against a set of target waveforms constructed here, produces match values better than 96.5% for systems heavier than $80M_\odot$, while with the old model this limit on the total mass is $115M_\odot$. The updated model is validated against an independent eccentric waveform family (TEOBResumS-Dali) for an initial eccentricity ($e_0$), mass ratio ($q$) and mean anomaly ($l_0$) in the range $0\lesssim e_0\lesssim0.3$, $1\lesssim q\lesssim3$ and $-\pi\leq l_0\leq\pi$, respectively. Further, an alternate model including the effect of higher order modes is also provided. Finally, while our model assumes nonspinning components, we show that it could also be used for systems with component spin vectors (anti-) aligned w.r.t. the orbital angular momentum and small spin magnitudes. | [
{
"version": "v1",
"created": "Mon, 16 Sep 2024 19:13:32 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 17:01:49 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Manna",
"Pratul",
""
],
[
"RoyChowdhury",
"Tamal",
""
],
[
"Mishra",
"Chandra Kant",
""
]
]
|
2409.11823 | Mehdi Heydari Shahna | Mehdi Heydari Shahna, Pauli Mustalahti, Jouni Mattila | Robust Sensor-Limited Control with Safe Input-Output Constraints for Hydraulic In-Wheel Motor Drive Mobility Systems | This work has been submitted for possible publication in the Elsevier | null | 10.1016/j.conengprac.2025.106459 | null | eess.SY cs.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In-wheel drive (IWD) systems enhance the responsiveness, traction, and maintenance efficiency of vehicles by enabling each wheel to operate independently. This paper proposes a novel robust torque-observed valve-based control (RTOVC) framework to address velocity tracking in hydraulic IWDs that actuate heavy-duty wheeled mobile robots (HWMRs), considering such challenges as wheel slippages, sensor limitations, rough terrains, and modeling uncertainties. To overcome the sensor-dependent control systems associated with the closed-loop torque/pressure in hydraulic IWD-actuated HWMRs, a robust observer network based on an adaptive barrier Lyapunov function (BLF) is proposed to estimate the required in-wheel motor torque to track the velocity references. Then, another adaptive BLF for valve control signals is employed to modulate the hydraulic fluid to generate the estimated torque for each IWD. The RTOVC strategy ensures user-defined safety within the logarithmic BLF framework by constraining the valve control signal, actual velocity, velocity tracking error, and torque of each hydraulic IWD in an HWMR to avoid exceeding specified limits. Despite its safety constraints, external disturbances, and modeling uncertainties, robustness and uniformly exponential stability of the RTOVC-applied hydraulic IWD mechanism are ensured in HWMRs. Experimental investigations using a 6,500-kg HWMR, actuated by four independent IWDs under intense disturbances and safety-defined constraints, validate the performance of the RTOVC. | [
{
"version": "v1",
"created": "Wed, 18 Sep 2024 09:16:03 GMT"
},
{
"version": "v2",
"created": "Wed, 04 Dec 2024 16:30:07 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Shahna",
"Mehdi Heydari",
""
],
[
"Mustalahti",
"Pauli",
""
],
[
"Mattila",
"Jouni",
""
]
]
|
2409.12546 | Divya Khurana | Divya Khurana | On orthogonality preserving and reversing operators | null | null | null | null | math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study approximately orthogonality (in the sense of Dragomir) preserving and reversing operators. We obtain a complete characterization of approximate orthogonality preserving and reversing operators for a class of operators. Using this characterization, we show that for some orthogonality notations, an operator defined from a finite-dimensional Banach space to a normed linear space is approximately orthogonality preserving/reversing if and only if it is an injective operator. This result implies that for some orthogonality notations, any operator defined from an $n$-dimensional Banach space to another $n$-dimensional Banach space is approximately orthogonality preserving/reversing if and only if it is a scalar multiple of an $\varepsilon$-isometry. We show that any $\varepsilon$-isometry and maps close to $\varepsilon$-isometries defined from a normed linear space to another normed linear space are approximately orthogonality preserving/reversing for some orthogonality notations. We also study the locally approximate orthogonality preserving and reversing operators defined on some finite-dimensional Banach spaces. | [
{
"version": "v1",
"created": "Thu, 19 Sep 2024 08:07:40 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Sep 2024 05:23:58 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 10:20:58 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Khurana",
"Divya",
""
]
]
|
2409.12850 | Chun-Yi Wang | Chun-Yi Wang and Fu-Hau Hsu | USBIPS Framework: Protecting Hosts from Malicious USB Peripherals | Accepted 23 June 2025 by Computer Standards & Interfaces/Elsevier B.V | null | 10.1016/j.csi.2025.104040 | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Universal Serial Bus (USB)-based attacks have increased in complexity in recent years. Modern attacks incorporate a wide range of attack vectors, from social engineering to signal injection. The security community is addressing these challenges using a growing set of fragmented defenses. Regardless of the vector of a USB-based attack, the most important risks concerning most people and enterprises are service crashes and data loss. The host OS manages USB peripherals, and malicious USB peripherals, such as those infected with BadUSB, can crash a service or steal data from the OS. Although USB firewalls have been proposed to thwart malicious USB peripherals, such as USBFilter and USBGuard, their effect is limited for preventing real-world intrusions. This paper focuses on building a security framework called USBIPS within Windows OSs to defend against malicious USB peripherals. This includes major efforts to explore the nature of malicious behavior and achieve persistent protection from USB-based intrusions. Herein, we first introduce an allowlisting-based method for USB access control. We then present a behavior-based detection mechanism focusing on attacks integrated into USB peripherals. Finally, we propose a novel approach that combines cross-layer methods to build the first generic security framework that thwarts USB-based intrusions. Within a centralized threat analysis framework, the approach provides persistent protection and may detect unknown malicious behavior. By addressing key security and performance challenges, these efforts help modern OSs against attacks from untrusted USB peripherals. | [
{
"version": "v1",
"created": "Thu, 19 Sep 2024 15:48:52 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 05:23:52 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Wang",
"Chun-Yi",
""
],
[
"Hsu",
"Fu-Hau",
""
]
]
|
2409.14052 | Mayank Deora | Mayank Deora, Pinakpani Pal | An average case efficient algorithm for solving two-variable linear Diophantine equations | null | null | null | null | cs.CR cs.DS math.NT | http://creativecommons.org/licenses/by/4.0/ | Solving two-variable linear Diophantine equations has applications in many cryptographic protocols such as RSA and Elliptic curve cryptography. The Extended Euclid's algorithm is the most widely used algorithm to solve these equations. We revisit two algorithms to solve two-variable linear Diophantine equations. We write the iterative version of one of the revisited algorithms. For another, we do a fine-grained analysis of the number of recursive calls and arrive at a periodic function that represents the number of recursive calls. We find the period and use it to derive an accurate closed-form expression for the average number of recursive calls incurred by that algorithm. We find multiple loose upper bounds on the average number of recursive calls in different cases based on whether a solution exists or not. We find that for a fixed value of $a,b$ and a varying $c$, such that the equation $ax+by=c$ (where $a > b$) is solvable, we can find the solution in $O\left(\frac{\log b}{gcd(a,b)}\right)$ average number of recursions or steps. We computationally evaluate this bound as well as one more upper bound and compare them with the average number of recursive calls in Extended Euclid's algorithm on a number of random $512$-bit inputs. We observe that the average number of iterations in the analyzed algorithm decreases with an increase in $gcd(a,b)$. We propose an iterative version of the algorithm. We implement this algorithm and find that the average number of iterations by our algorithm is less than that of two existing algorithms. | [
{
"version": "v1",
"created": "Sat, 21 Sep 2024 07:51:12 GMT"
},
{
"version": "v2",
"created": "Tue, 29 Apr 2025 07:05:41 GMT"
},
{
"version": "v3",
"created": "Wed, 21 May 2025 07:04:11 GMT"
},
{
"version": "v4",
"created": "Tue, 27 May 2025 07:31:15 GMT"
},
{
"version": "v5",
"created": "Thu, 03 Jul 2025 07:08:25 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Deora",
"Mayank",
""
],
[
"Pal",
"Pinakpani",
""
]
]
|
2409.14450 | Yury Budkov | P.E. Brandyshev and Yu.A. Budkov | Theory of Casimir Forces: A Unified Approach Using Finite-Temperature Field Theory | This manuscript will be published in The Journal of Mathematical Physics | null | null | null | cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | We present a quantum theory of Casimir forces between perfect electrical conductors, based on quantum electrodynamics and quantum statistical physics. This theory utilizes Kapusta's finite-temperature quantum field theory, combined with the Faddeev-Popov ghost formalism. This approach allows us to calculate Casimir forces at finite temperatures, providing both previously known and new physical insights from a unified perspective. Furthermore, our method enables us to compute the stress tensor associated with Casimir forces, in accordance with the Helmholtz free energy of an equilibrium quantum electromagnetic field. Using this methodology, we calculate the excess surface tension on the walls of slit-like pores due to the Casimir effect. | [
{
"version": "v1",
"created": "Sun, 22 Sep 2024 14:01:31 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2024 07:22:46 GMT"
},
{
"version": "v3",
"created": "Tue, 25 Feb 2025 18:02:17 GMT"
},
{
"version": "v4",
"created": "Tue, 18 Mar 2025 16:32:16 GMT"
},
{
"version": "v5",
"created": "Thu, 03 Jul 2025 11:05:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Brandyshev",
"P. E.",
""
],
[
"Budkov",
"Yu. A.",
""
]
]
|
2409.15582 | Alex Nguyen | Alex Nguyen, David J. Schwab, Vudtiwat Ngampruetikorn | Generalization vs. Specialization under Concept Shift | 8 pages, 3 figures | null | null | null | stat.ML cond-mat.dis-nn cond-mat.stat-mech cs.LG | http://creativecommons.org/licenses/by/4.0/ | Machine learning models are often brittle under distribution shift, i.e., when data distributions at test time differ from those during training. Understanding this failure mode is central to identifying and mitigating safety risks of mass adoption of machine learning. Here we analyze ridge regression under concept shift -- a form of distribution shift in which the input-label relationship changes at test time. We derive an exact expression for prediction risk in the thermodynamic limit. Our results reveal nontrivial effects of concept shift on generalization performance, including a phase transition between weak and strong concept shift regimes and nonmonotonic data dependence of test performance even when double descent is absent. Our theoretical results are in good agreement with experiments based on transformers pretrained to solve linear regression; under concept shift, too long context length can be detrimental to generalization performance of next token prediction. Finally, our experiments on MNIST and FashionMNIST suggest that this intriguing behavior is present also in classification problems. | [
{
"version": "v1",
"created": "Mon, 23 Sep 2024 22:30:28 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 05:05:13 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Nguyen",
"Alex",
""
],
[
"Schwab",
"David J.",
""
],
[
"Ngampruetikorn",
"Vudtiwat",
""
]
]
|
2409.17015 | Tilo Strutz | Tilo Strutz and Nico Schreiber | Investigations on Algorithm Selection for Interval-Based Coding Methods | null | Multimedia Tools and Applications, Published online 03 July 2025 | 10.1007/s11042-025-20971-3 | null | cs.IT cs.DS math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is a class of entropy-coding methods which do not substitute symbols by code words (such as Huffman coding), but operate on intervals or ranges. This class includes three prominent members: conventional arithmetic coding, range coding, and coding based on asymmetric numeral systems. To determine the correct symbol in the decoder, each of these methods requires the comparison of a state variable with subinterval boundaries. In adaptive operation, considering varying symbol statistics, an array of interval boundaries must additionally be kept up to date. The larger the symbol alphabet, the more time-consuming both the search for the correct subinterval and the updating of interval borders become.
Detailed pseudo-code is used to discuss different approaches to speed up the symbol search in the decoder and the adaptation of the array of interval borders, both depending on the chosen alphabet size. It is shown that reducing the $\mathcal{O}$-complexity does not lead to an acceleration in practical implementations if the alphabet size is too small. In adaptive compression mode, the binary indexing method proves to be superior when considering the overall processing time. Although the symbol search (in the decoder) takes longer than with other algorithms, the faster updating of the array of interval borders more than compensates for this disadvantage. A variant of the binary indexing method is proposed, which is more flexible and has a partially lower complexity than the original approach. | [
{
"version": "v1",
"created": "Wed, 25 Sep 2024 15:17:40 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Strutz",
"Tilo",
""
],
[
"Schreiber",
"Nico",
""
]
]
|
2409.18624 | Alfredo Ibias | Alfredo Ibias, Hector Antona, Guillem Ramirez-Miranda, Enric Guinovart, Eduard Alarcon | Unsupervised Cognition | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Unsupervised learning methods have a soft inspiration in cognition models. To this day, the most successful unsupervised learning methods revolve around clustering samples in a mathematical space. In this paper we propose a primitive-based, unsupervised learning approach for decision-making inspired by a novel cognition framework. This representation-centric approach models the input space constructively as a distributed hierarchical structure in an input-agnostic way. We compared our approach with both current state-of-the-art unsupervised learning classification, with current state-of-the-art small and incomplete datasets classification, and with current state-of-the-art cancer type classification. We show how our proposal outperforms previous state-of-the-art. We also evaluate some cognition-like properties of our proposal where it not only outperforms the compared algorithms (even supervised learning ones), but it also shows a different, more cognition-like, behaviour. | [
{
"version": "v1",
"created": "Fri, 27 Sep 2024 10:50:49 GMT"
},
{
"version": "v2",
"created": "Thu, 07 Nov 2024 14:36:04 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 08:53:17 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Ibias",
"Alfredo",
""
],
[
"Antona",
"Hector",
""
],
[
"Ramirez-Miranda",
"Guillem",
""
],
[
"Guinovart",
"Enric",
""
],
[
"Alarcon",
"Eduard",
""
]
]
|
2409.18954 | Sandra Brandstetter | Sandra Brandstetter, Carl Heintze, Paul Hill, Philipp M. Preiss, Maciej Ga{\l}ka and Selim Jochim | Magnifying the Wave Function of Interacting Fermionic Atoms | 7 pages, 4 figures | null | 10.1103/wdjr-m2hg | null | cond-mat.quant-gas quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding many body systems is a key challenge in physics. Single atom resolved imaging techniques have unlocked access to microscopic correlations in ultracold quantum gases. However they cannot be used when the relevant length scales are obscured by the resolution of the detection technique. We present a matterwave magnification scheme, based on evolutions in optical potentials, tailored to magnify the wave function of atoms, such that all length scales can be resolved. To showcase this method, we image atoms in the strongly interacting regime, establishing a new way to characterize correlated systems. | [
{
"version": "v1",
"created": "Fri, 27 Sep 2024 17:57:40 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 16:04:28 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Brandstetter",
"Sandra",
""
],
[
"Heintze",
"Carl",
""
],
[
"Hill",
"Paul",
""
],
[
"Preiss",
"Philipp M.",
""
],
[
"Gałka",
"Maciej",
""
],
[
"Jochim",
"Selim",
""
]
]
|
2409.19031 | Johann (Jan) Rafelski | Johann Rafelski, Jeremiah Birrell, Christopher Grayson, Andrew Steinmetz, and Cheng Tao Yang | Quarks to Cosmos: Particles and Plasma in Cosmological evolution | 189 pages review, includes 90 figure frames. Incluses key results from four PhD thesis and two dozen research papers. See Appendix A for these details. Reformatted for new widescreen of The European Physical Journal - Special Topics (EPJ ST) format with a few typos corrected, | Eur. Phys. J. Spec. Top. (2025) | 10.1140/epjs/s11734-025-01470-w | null | hep-ph astro-ph.CO nucl-th physics.plasm-ph | http://creativecommons.org/licenses/by/4.0/ | We describe in the context of the particle physics (PP) standard model (SM) `PP-SM' the understanding of the primordial properties and composition of the Universe in the temperature range $130\GeV>T>20\keV$. The Universe evolution is described using FLRW cosmology. We present a global view on particle content across time and describe the different evolution eras using deceleration parameter $q$. We follow the arrow of time in the expanding and cooling Universe: After the PP-SM heavies $(t, h, W, Z)$ diminish in abundance below $T\simeq 50\GeV$, the PP-SM plasma in the Universe is governed by the strongly interacting Quark-Gluon content. Once the temperature drops below $T\simeq 150\MeV$, quarks and gluons hadronize into strongly interacting matter particles. Rapid disappearance of baryonic antimatter completes at $T_\mathrm{B}=38.2\MeV$. We study the ensuing disappearance of strangeness and mesons in general. We show that the different eras defined by particle populations are barely separated from each other with abundance of muons fading out just prior to $T=\mathcal{O}(2.5)\MeV$, the era of emergence of the free-streaming neutrinos. We discuss the two relevant fundamental constants controlling the decoupling of neutrinos. We subsequently follow the primordial Universe as it passes through the hot dense electron-positron plasma epoch. The high density of positron antimatter disappears near $T=20.3\keV$: Nuclear reactions occur in the presence of a highly mobile and relatively strongly interacting electron-positron plasma phase. We apply plasma theory methods to describe the strong screening effects between heavy dust particle (nucleons). We analyze the paramagnetic characteristics of the electron-positron plasma when exposed to an external primordial magnetic field. | [
{
"version": "v1",
"created": "Thu, 26 Sep 2024 21:24:05 GMT"
},
{
"version": "v2",
"created": "Tue, 14 Jan 2025 23:15:28 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Rafelski",
"Johann",
""
],
[
"Birrell",
"Jeremiah",
""
],
[
"Grayson",
"Christopher",
""
],
[
"Steinmetz",
"Andrew",
""
],
[
"Yang",
"Cheng Tao",
""
]
]
|
2409.19303 | Ken'ichi Yoshida | Ken'ichi Yoshida | Links in the spherical 3-manifold obtained from the quaternion group and their lifts | 20 pages, 19 figures | null | null | null | math.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that there are infinitely many triples of non-isotopic hyperbolic links in the lens space $L(4,1)$ such that the three lifts of each triple in $S^{3}$ are isotopic. They are obtained as the lifts of links in $S^{3} / Q_{8}$ by double covers, where $Q_{8}$ is the quaternion group. To construct specific examples, we introduce a diagram of a link in $S^{3} / Q_{8}$ obtained by projecting to a square. The diagrams of isotopic links are connected by Reidemeister-type moves. | [
{
"version": "v1",
"created": "Sat, 28 Sep 2024 10:20:50 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 12:42:30 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yoshida",
"Ken'ichi",
""
]
]
|
2410.00903 | Kentaro Nakamura | Kosuke Imai, Kentaro Nakamura | Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments | null | null | null | null | stat.AP cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | In this paper, we demonstrate how to enhance the validity of causal inference with unstructured high-dimensional treatments like texts, by leveraging the power of generative Artificial Intelligence (GenAI). Specifically, we propose to use a deep generative model such as large language models (LLMs) to efficiently generate treatments and use their internal representation for subsequent causal effect estimation. We show that the knowledge of this true internal representation helps disentangle the treatment features of interest, such as specific sentiments and certain topics, from other possibly unknown confounding features. Unlike existing methods, the proposed GenAI-Powered Inference (GPI) methodology eliminates the need to learn causal representation from the data, and hence produces more accurate and efficient estimates. We formally establish the conditions required for the nonparametric identification of the average treatment effect, propose an estimation strategy that avoids the violation of the overlap assumption, and derive the asymptotic properties of the proposed estimator through the application of double machine learning. Finally, using an instrumental variables approach, we extend the proposed methodology to the settings in which the treatment feature is based on human perception. The proposed GPI methodology is also applicable to text reuse where an LLM is used to regenerate existing texts. We conduct simulation and empirical studies, using the generated text data from an open-source LLM, Llama 3, to illustrate the advantages of our estimator over state-of-the-art causal representation learning algorithms. | [
{
"version": "v1",
"created": "Tue, 01 Oct 2024 17:46:21 GMT"
},
{
"version": "v2",
"created": "Fri, 22 Nov 2024 16:31:25 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 03:31:52 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Imai",
"Kosuke",
""
],
[
"Nakamura",
"Kentaro",
""
]
]
|
2410.01006 | Jayasmita Behera | Salil Bedkihal, Jayasmita Behera, and Malay Bandyopadhyay | A review on Aharonov-Bohm quantum machines: Thermoelectric heat engines and diodes | null | J. Phys.: Condens. Matter 37 163001 (2025) | 10.1088/1361-648X/adb921 | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | The study of heat-to-work conversion has gained significant attention in recent years, highlighting the potential of nanoscale systems to achieve energy conversion in steady-state devices without any macroscopic moving parts. This review examines the theoretical frameworks governing the steady-state flows of quantum particles like electrons, photons, and phonons within various mesoscopic or nanoscale devices, such as thermoelectric heat engines in the context of quantum dot Aharonov-Bohm (AB) interferometric configurations. Quantum interference effects hold great promise for enhancing the thermoelectric transport properties of such quantum devices by allowing more precise control over energy levels and transport pathways. Driven quantum dot AB networks can maintain quantum coherence and provide precise experimental control. Unlike bulk systems, nanoscale systems like quantum dots reveal distinct quantum interference phenomena, including sharp features in transmission spectra and Fano resonances. This review highlights the distinction between optimization methods that produce boxcar functions and coherent control methods that result in complex interference patterns. It reveals that the effective design of thermoelectric heat engines requires careful tailoring of quantum interference and the magnetic field-induced effects to enhance performance. We emphasize how magnetic fields can change the bounds of power or efficiency. These machines with broken time-reversal symmetry provide insights into directional dependencies and asymmetries in quantum transport. We offer a thorough overview of past and current research on quantum thermoelectric heat engines using the AB effect and present a detailed review of three-terminal AB heat engines, where broken time-reversal symmetry can induce a coherent diode effect. We cover bounds on power and efficiency in systems with broken time-reversal symmetry. | [
{
"version": "v1",
"created": "Tue, 01 Oct 2024 18:55:16 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Bedkihal",
"Salil",
""
],
[
"Behera",
"Jayasmita",
""
],
[
"Bandyopadhyay",
"Malay",
""
]
]
|
2410.03002 | T. M. Dunster | T. M. Dunster | Simplified uniform asymptotic expansions for associated Legendre and conical functions | null | null | null | null | math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Asymptotic expansions are derived for associated Legendre functions of degree $\nu$ and order $\mu$, where one or the other of the parameters is large. The expansions are uniformly valid for unbounded real and complex values of the argument $z$, including the singularity $z=1$. The cases where $\nu+\frac12$ and $\mu$ are real or purely imaginary are included, which covers conical functions. The approximations involve either exponential or modified Bessel functions, along with slowly varying coefficient functions. The coefficients of the new asymptotic expansions are simple and readily obtained explicitly, allowing for computation to a high degree of accuracy. The results are constructed and rigorously established by employing certain Liouville-Green type expansions where the coefficients appear in the exponent of an exponential function. | [
{
"version": "v1",
"created": "Thu, 03 Oct 2024 21:23:03 GMT"
},
{
"version": "v2",
"created": "Mon, 14 Oct 2024 02:27:04 GMT"
},
{
"version": "v3",
"created": "Wed, 02 Jul 2025 22:25:42 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Dunster",
"T. M.",
""
]
]
|
2410.04813 | Arnaud Allera | Arnaud Allera, Thomas D. Swinburne, Alexandra M. Goryaeva, Baptiste Bienvenu, Fabienne Ribeiro, Michel Perez, Mihai-Cosmin Marinica, David Rodney | Activation entropy of dislocation glide in body-centered cubic metals from atomistic simulations | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The activation entropy of dislocation glide, a key process controlling the strength of many metals, is often assumed to be constant or linked to enthalpy through the empirical Meyer-Neldel law-both of which are simplified approximations. In this study, we take a more direct approach by calculating the activation Gibbs energy for kink-pair nucleation on screw dislocations of two body-centered cubic metals, iron and tungsten. To ensure reliability, we develop machine learning interatomic potentials for both metals, carefully trained on dislocation data from density functional theory. Our findings reveal that dislocations undergo harmonic transitions between Peierls valleys, with an activation entropy that remains largely constant, regardless of temperature or applied stress. We use these results to parameterize a thermally-activated model of yield stress, which consistently matches experimental data in both iron and tungsten. Our work challenges recent studies using classical potentials, which report highly varying activation entropies, and suggests that simulations relying on classical potentials-widely used in materials modeling-could be significantly influenced by overestimated entropic effects. | [
{
"version": "v1",
"created": "Mon, 07 Oct 2024 07:53:40 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 14:04:14 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Allera",
"Arnaud",
""
],
[
"Swinburne",
"Thomas D.",
""
],
[
"Goryaeva",
"Alexandra M.",
""
],
[
"Bienvenu",
"Baptiste",
""
],
[
"Ribeiro",
"Fabienne",
""
],
[
"Perez",
"Michel",
""
],
[
"Marinica",
"Mihai-Cosmin",
""
],
[
"Rodney",
"David",
""
]
]
|
2410.05451 | Sizhe Chen | Sizhe Chen, Arman Zharmagambetov, Saeed Mahloujifar, Kamalika Chaudhuri, David Wagner, Chuan Guo | SecAlign: Defending Against Prompt Injection with Preference Optimization | ACM CCS 2025. Key words: prompt injection defense, LLM security, LLM-integrated applications | null | 10.1145/3719027.3744836 | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) are becoming increasingly prevalent in modern software systems, interfacing between the user and the Internet to assist with tasks that require advanced language understanding. To accomplish these tasks, the LLM often uses external data sources such as user documents, web retrieval, results from API calls, etc. This opens up new avenues for attackers to manipulate the LLM via prompt injection. Adversarial prompts can be injected into external data sources to override the system's intended instruction and instead execute a malicious instruction. To mitigate this vulnerability, we propose a new defense called SecAlign based on the technique of preference optimization. Our defense first constructs a preference dataset with prompt-injected inputs, secure outputs (ones that respond to the legitimate instruction), and insecure outputs (ones that respond to the injection). We then perform preference optimization on this dataset to teach the LLM to prefer the secure output over the insecure one. This provides the first known method that reduces the success rates of various prompt injections to <10%, even against attacks much more sophisticated than ones seen during training. This indicates our defense generalizes well against unknown and yet-to-come attacks. Also, SecAlign models are still practical with similar utility to the one before defensive training in our evaluations. Our code is at https://github.com/facebookresearch/SecAlign | [
{
"version": "v1",
"created": "Mon, 07 Oct 2024 19:34:35 GMT"
},
{
"version": "v2",
"created": "Mon, 13 Jan 2025 18:45:57 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 05:45:41 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Chen",
"Sizhe",
""
],
[
"Zharmagambetov",
"Arman",
""
],
[
"Mahloujifar",
"Saeed",
""
],
[
"Chaudhuri",
"Kamalika",
""
],
[
"Wagner",
"David",
""
],
[
"Guo",
"Chuan",
""
]
]
|
2410.06579 | Paul Jolissaint | Paul Jolissaint | Non Kazhdan's groups and a new approximation property for tracial von Neumann algebras | A serious gap in the proof of the main theorem has been corrected due to the remarks of the referee; an example of a diffuse, tracial von Neumann algebra which has neither property (T) nor the WMAP is given | null | null | null | math.OA math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We define a new approximation property for tracial von Neumann algebras, called \textit{weakly mixing approximation property} which, for discrete groups and II$_1$ factors, is equivalent to the negation of Kazhdan's property (T). | [
{
"version": "v1",
"created": "Wed, 09 Oct 2024 06:23:35 GMT"
},
{
"version": "v2",
"created": "Mon, 21 Oct 2024 07:26:01 GMT"
},
{
"version": "v3",
"created": "Sun, 03 Nov 2024 21:57:38 GMT"
},
{
"version": "v4",
"created": "Mon, 10 Feb 2025 08:57:10 GMT"
},
{
"version": "v5",
"created": "Wed, 14 May 2025 09:04:53 GMT"
},
{
"version": "v6",
"created": "Wed, 02 Jul 2025 20:23:46 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Jolissaint",
"Paul",
""
]
]
|
2410.07423 | Tamar Friedmann | Tamar Friedmann | The construction of a class of presentations for Specht modules | 24 pages; this article generalizes Theorem 2.4 or arXiv:1710.00376, Theorem 1.1 of arXiv:2303.00587, and Theorem 3.5 of arXiv:1803.10918; v2: minor revisions, version to appear in The Electronic Journal of Combinatorics | The Electronic Journal of Combinatorics 32(1) (2025) #P1.46 | null | null | math.CO math.RT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We build on the methods introduced by Friedmann, Hanlon, Stanley, and Wachs, and further developed by Brauner and Friedmann, to construct additional classes of presentations of Specht modules. We obtain these presentations by defining a linear operator which is a symmetrized sum of dual Garnir relations on the space of column tabloids. Our presentations apply to the vast majority of shapes of Specht modules. | [
{
"version": "v1",
"created": "Wed, 09 Oct 2024 20:39:36 GMT"
},
{
"version": "v2",
"created": "Sun, 23 Feb 2025 16:08:12 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Friedmann",
"Tamar",
""
]
]
|
2410.08500 | Yunpeng Gao | Yunpeng Gao, Zhigang Wang, Linglin Jing, Dong Wang, Xuelong Li, Bin Zhao | Aerial Vision-and-Language Navigation via Semantic-Topo-Metric Representation Guided LLM Reasoning | null | null | null | null | cs.RO cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aerial Vision-and-Language Navigation (VLN) is a novel task enabling Unmanned Aerial Vehicles (UAVs) to navigate in outdoor environments through natural language instructions and visual cues. It remains challenging due to the complex spatial relationships in outdoor aerial scenes. In this paper, we propose an end-to-end zero-shot framework for aerial VLN tasks, where the large language model (LLM) is introduced as our agent for action prediction. Specifically, we develop a novel Semantic-Topo-Metric Representation (STMR) to enhance the spatial reasoning ability of LLMs. This is achieved by extracting and projecting instruction-related semantic masks of landmarks into a top-down map that contains the location information of surrounding landmarks. Further, this map is transformed into a matrix representation with distance metrics as the text prompt to the LLM, for action prediction according to the instruction. Experiments conducted in real and simulation environments have successfully proved the effectiveness and robustness of our method, achieving 15.9% and 12.5% improvements (absolute) in Oracle Success Rate (OSR) on AerialVLN-S dataset. | [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 03:54:48 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 11:55:43 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Gao",
"Yunpeng",
""
],
[
"Wang",
"Zhigang",
""
],
[
"Jing",
"Linglin",
""
],
[
"Wang",
"Dong",
""
],
[
"Li",
"Xuelong",
""
],
[
"Zhao",
"Bin",
""
]
]
|
2410.08723 | Yunchao Wang | Yunchao Wang, Guodao Sun, Zihang Fu, Ronghua Liang | Human-Computer Interaction and Visualization in Natural Language Generation Models: Applications, Challenges, and Opportunities | The article has been accepted by Frontiers of Computer Science (FCS), with the DOI: {10.1007/s11704-025-50356-6} | null | 10.1007/s11704-025-50356-6 | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Natural language generation (NLG) models have emerged as a focal point of research within natural language processing (NLP), exhibiting remarkable performance in tasks such as text composition and dialogue generation. However, their intricate architectures and extensive model parameters pose significant challenges to interpretability, limiting their applicability in high-stakes decision-making scenarios. To address this issue, human-computer interaction (HCI) and visualization techniques offer promising avenues to enhance the transparency and usability of NLG models by making their decision-making processes more interpretable. In this paper, we provide a comprehensive investigation into the roles, limitations, and impact of HCI and visualization in facilitating human understanding and control over NLG systems. We introduce a taxonomy of interaction methods and visualization techniques, categorizing three major research domains and their corresponding six key tasks in the application of NLG models. Finally, we summarize the shortcomings in the existing work and investigate the key challenges and emerging opportunities in the era of large language models (LLMs). | [
{
"version": "v1",
"created": "Fri, 11 Oct 2024 11:23:26 GMT"
},
{
"version": "v2",
"created": "Fri, 28 Mar 2025 04:38:44 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 05:25:23 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Wang",
"Yunchao",
""
],
[
"Sun",
"Guodao",
""
],
[
"Fu",
"Zihang",
""
],
[
"Liang",
"Ronghua",
""
]
]
|
2410.10198 | Jinxing Yang | Yanru Chen, Suijie Wang, Jinxing Yang, Chengdong Zhao | Regions of Level $\ell$ of Catalan/Semiorder-Type Arrangements | 38 pages,16 figures | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In 1996, Stanley extended the classical Catalan arrangement and semiorder arrangement, which are called the Catalan-type arrangement $\mathcal{C}_{n,A}$ and the semiorder-type arrangement $\mathcal{C}_{n,A}^*$ in this paper. By establishing a labeled Dyck path model for the regions of $\mathcal{C}_{n,A}$ and $\mathcal{C}_{n,A}^*$, this paper explores several enumerative problems related to the number of regions of level $\ell$, denoted as $r_{\ell}(\mathcal{C}_{n,A})$ and $r_{\ell}(\mathcal{C}_{n,A}^*)$, which includes:
(1) proving a Stirling convolution relation between $r_{\ell}(\mathcal{C}_{n,A})$ and $r_{\ell}(\mathcal{C}_{n,A}^*)$, refining a result by Stanley and Postnikov;
(2) showing that the sequences$\left(r_\ell{(\mathcal{C}_{n,A})}\right)_{n\geq 0}$ and $(r_\ell {(\mathcal{C}_{n,A}^*)})_{n\geq 0}$ exhibit properties of binomial type in the sense of Rota;
(3) establishing the transformational significance of $r_{\ell}(\mathcal{C}_{n,A})$ and $r_{\ell}(\mathcal{C}_{n,A}^*)$ under Stanley's ESA framework: they can be viewed as transition matrices from binomial coefficients to their characteristic polynomials respectively.
Further, we present two applications of the theories and methods: first, we provide a hyperplane arrangement counting interpretation of the two-parameter generalization of Fuss--Catalan numbers, which is closely related to the number of regions of level $\ell$ in the $m$-Catalan arrangement. Second, using labeled Dyck paths to depict the number of regions in the $m$-Catalan arrangement, we algorithmically provide the inverse mapping of the Fu, Wang, and Zhu mapping. | [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 06:35:38 GMT"
},
{
"version": "v2",
"created": "Tue, 15 Oct 2024 14:07:50 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 08:09:24 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Chen",
"Yanru",
""
],
[
"Wang",
"Suijie",
""
],
[
"Yang",
"Jinxing",
""
],
[
"Zhao",
"Chengdong",
""
]
]
|
2410.10530 | Nicholas Kr\"amer | Nicholas Kr\"amer | Adaptive Probabilistic ODE Solvers Without Adaptive Memory Requirements | null | null | null | null | math.NA cs.LG cs.NA stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Despite substantial progress in recent years, probabilistic solvers with adaptive step sizes can still not solve memory-demanding differential equations -- unless we care only about a single point in time (which is far too restrictive; we want the whole time series). Counterintuitively, the culprit is the adaptivity itself: Its unpredictable memory demands easily exceed our machine's capabilities, making our simulations fail unexpectedly and without warning. Still, dropping adaptivity would abandon years of progress, which can't be the answer. In this work, we solve this conundrum. We develop an adaptive probabilistic solver with fixed memory demands building on recent developments in robust state estimation. Switching to our method (i) eliminates memory issues for long time series, (ii) accelerates simulations by orders of magnitude through unlocking just-in-time compilation, and (iii) makes adaptive probabilistic solvers compatible with scientific computing in JAX. | [
{
"version": "v1",
"created": "Mon, 14 Oct 2024 14:10:47 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 12:07:20 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Krämer",
"Nicholas",
""
]
]
|
2410.12532 | Dingkang Yang | Dingkang Yang, Jinjie Wei, Mingcheng Li, Jiyao Liu, Lihao Liu, Ming Hu, Junjun He, Yakun Ju, Wei Zhou, Yang Liu, Lihua Zhang | MedAide: Information Fusion and Anatomy of Medical Intents via LLM-based Agent Collaboration | LLM-based Multi-Agent Collaboration for Medical Applications | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In healthcare intelligence, the ability to fuse heterogeneous, multi-intent information from diverse clinical sources is fundamental to building reliable decision-making systems. Large Language Model (LLM)-driven information interaction systems currently showing potential promise in the healthcare domain. Nevertheless, they often suffer from information redundancy and coupling when dealing with complex medical intents, leading to severe hallucinations and performance bottlenecks. To this end, we propose MedAide, an LLM-based medical multi-agent collaboration framework designed to enable intent-aware information fusion and coordinated reasoning across specialized healthcare domains. Specifically, we introduce a regularization-guided module that combines syntactic constraints with retrieval augmented generation to decompose complex queries into structured representations, facilitating fine-grained clinical information fusion and intent resolution. Additionally, a dynamic intent prototype matching module is proposed to utilize dynamic prototype representation with a semantic similarity matching mechanism to achieve adaptive recognition and updating of the agent's intent in multi-round healthcare dialogues. Ultimately, we design a rotation agent collaboration mechanism that introduces dynamic role rotation and decision-level information fusion across specialized medical agents. Extensive experiments are conducted on four medical benchmarks with composite intents. Experimental results from automated metrics and expert doctor evaluations show that MedAide outperforms current LLMs and improves their medical proficiency and strategic reasoning. | [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 13:10:27 GMT"
},
{
"version": "v2",
"created": "Thu, 17 Oct 2024 09:22:41 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 13:02:30 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yang",
"Dingkang",
""
],
[
"Wei",
"Jinjie",
""
],
[
"Li",
"Mingcheng",
""
],
[
"Liu",
"Jiyao",
""
],
[
"Liu",
"Lihao",
""
],
[
"Hu",
"Ming",
""
],
[
"He",
"Junjun",
""
],
[
"Ju",
"Yakun",
""
],
[
"Zhou",
"Wei",
""
],
[
"Liu",
"Yang",
""
],
[
"Zhang",
"Lihua",
""
]
]
|
2410.12537 | Cosimo Gregucci | Cosimo Gregucci, Bo Xiong, Daniel Hernandez, Lorenzo Loconte, Pasquale Minervini, Steffen Staab, Antonio Vergari | Is Complex Query Answering Really Complex? | ICML 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Complex query answering (CQA) on knowledge graphs (KGs) is gaining momentum as a challenging reasoning task. In this paper, we show that the current benchmarks for CQA might not be as complex as we think, as the way they are built distorts our perception of progress in this field. For example, we find that in these benchmarks, most queries (up to 98% for some query types) can be reduced to simpler problems, e.g., link prediction, where only one link needs to be predicted. The performance of state-of-the-art CQA models decreases significantly when such models are evaluated on queries that cannot be reduced to easier types. Thus, we propose a set of more challenging benchmarks composed of queries that require models to reason over multiple hops and better reflect the construction of real-world KGs. In a systematic empirical investigation, the new benchmarks show that current methods leave much to be desired from current CQA methods. | [
{
"version": "v1",
"created": "Wed, 16 Oct 2024 13:19:03 GMT"
},
{
"version": "v2",
"created": "Tue, 18 Feb 2025 12:40:13 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 08:36:22 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Gregucci",
"Cosimo",
""
],
[
"Xiong",
"Bo",
""
],
[
"Hernandez",
"Daniel",
""
],
[
"Loconte",
"Lorenzo",
""
],
[
"Minervini",
"Pasquale",
""
],
[
"Staab",
"Steffen",
""
],
[
"Vergari",
"Antonio",
""
]
]
|
2410.13808 | Ruibo Chen | Ruibo Chen, Yihan Wu, Junfeng Guo, Heng Huang | De-mark: Watermark Removal in Large Language Models | ICML 2025 | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Watermarking techniques offer a promising way to identify machine-generated content via embedding covert information into the contents generated from language models (LMs). However, the robustness of the watermarking schemes has not been well explored. In this paper, we present De-mark, an advanced framework designed to remove n-gram-based watermarks effectively. Our method utilizes a novel querying strategy, termed random selection probing, which aids in assessing the strength of the watermark and identifying the red-green list within the n-gram watermark. Experiments on popular LMs, such as Llama3 and ChatGPT, demonstrate the efficiency and effectiveness of De-mark in watermark removal and exploitation tasks. | [
{
"version": "v1",
"created": "Thu, 17 Oct 2024 17:42:10 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 20:37:50 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Chen",
"Ruibo",
""
],
[
"Wu",
"Yihan",
""
],
[
"Guo",
"Junfeng",
""
],
[
"Huang",
"Heng",
""
]
]
|
2410.16236 | Yuxuan Cai | Yuxuan Cai, Jiangning Zhang, Haoyang He, Xinwei He, Ao Tong, Zhenye Gan, Chengjie Wang, Zhucun Xue, Yong Liu, Xiang Bai | LLaVA-KD: A Framework of Distilling Multimodal Large Language Models | ICCV'25 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The success of Large Language Models (LLMs) has inspired the development of Multimodal Large Language Models (MLLMs) for unified understanding of vision and language. However, the increasing model size and computational complexity of large-scale MLLMs (l-MLLMs) limit their use in resource-constrained scenarios. Although small-scale MLLMs (s-MLLMs) are designed to reduce computational costs, they typically suffer from performance degradation. To mitigate this limitation, we propose a novel LLaVA-KD framework to transfer knowledge from l-MLLMs to s-MLLMs. Specifically, we introduce Multimodal Distillation (MDist) to transfer teacher model's robust representations across both visual and linguistic modalities, and Relation Distillation (RDist) to transfer teacher model's ability to capture visual token relationships. Additionally, we propose a three-stage training scheme to fully exploit the potential of the proposed distillation strategy: 1) Distilled Pre-Training to strengthen the alignment between visual-linguistic representations in s-MLLMs, 2) Supervised Fine-Tuning to equip the s-MLLMs with multimodal understanding capacity, and 3) Distilled Fine-Tuning to refine s-MLLM's knowledge. Our approach significantly improves s-MLLMs performance without altering the model architecture. Extensive experiments and ablation studies validate the effectiveness of each proposed component. Code will be available at https://github.com/Fantasyele/LLaVA-KD. | [
{
"version": "v1",
"created": "Mon, 21 Oct 2024 17:41:28 GMT"
},
{
"version": "v2",
"created": "Fri, 25 Oct 2024 06:19:13 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 06:59:56 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Cai",
"Yuxuan",
""
],
[
"Zhang",
"Jiangning",
""
],
[
"He",
"Haoyang",
""
],
[
"He",
"Xinwei",
""
],
[
"Tong",
"Ao",
""
],
[
"Gan",
"Zhenye",
""
],
[
"Wang",
"Chengjie",
""
],
[
"Xue",
"Zhucun",
""
],
[
"Liu",
"Yong",
""
],
[
"Bai",
"Xiang",
""
]
]
|
2410.16428 | Lantian Li Mr. | Wan Lin and Junhui Chen and Tianhao Wang and Zhenyu Zhou and Lantian Li and Dong Wang | Neural Scoring: A Refreshed End-to-End Approach for Speaker Recognition in Complex Conditions | null | null | null | null | cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Modern speaker verification systems primarily rely on speaker embeddings, followed by verification based on cosine similarity between the embedding vectors of the enrollment and test utterances. While effective, these methods struggle with multi-talker speech due to the unidentifiability of embedding vectors. In this paper, we propose Neural Scoring (NS), a refreshed end-to-end framework that directly estimates verification posterior probabilities without relying on test-side embeddings, making it more robust to complex conditions, e.g., with multiple talkers. To make the training of such an end-to-end model more efficient, we introduce a large-scale trial e2e training (LtE2E) strategy, where each test utterance pairs with a set of enrolled speakers, thus enabling the processing of large-scale verification trials per batch. Experiments on the VoxCeleb dataset demonstrate that NS consistently outperforms both the baseline and competitive methods across various conditions, achieving an overall 70.36% reduction in EER compared to the baseline. | [
{
"version": "v1",
"created": "Mon, 21 Oct 2024 18:48:54 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Jun 2025 23:35:07 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 16:36:28 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Lin",
"Wan",
""
],
[
"Chen",
"Junhui",
""
],
[
"Wang",
"Tianhao",
""
],
[
"Zhou",
"Zhenyu",
""
],
[
"Li",
"Lantian",
""
],
[
"Wang",
"Dong",
""
]
]
|
2410.18023 | Jayasmita Behera | Jayasmita Behera, Salil Bedkihal, Bijay Kumar Agarwalla, and Malay Bandyopadhyay | Thermoelectric performance of a minimally nonlinear voltage probe and voltage-temperature probe heat engine with broken time-reversal symmetry | null | null | null | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | We investigate the thermoelectric performance of minimally nonlinear irreversible heat engines with broken time-reversal symmetry (TRS), realized through voltage and voltage-temperature probe configurations. Our framework extends the Onsager relations by incorporating a nonlinear power dissipation term into the heat current. We derive and analyze analytical expressions for the efficiency at a given power and the efficiency at maximum power (EMP), expressed in terms of asymmetry parameters and generalized figures of merit. Our analysis reveals that the combined effects of broken TRS and nonlinear dissipation give rise to two universal bounds on the EMP that can surpass the Curzon-Ahlborn (CA) limit. Although these bounds share a similar analytical form, differences in Carnot efficiency and asymmetry parameters lead to distinct operational characteristics, as shown through numerical simulations. We consider a triple-quantum-dot Aharonov-Bohm heat engine incorporating either a voltage probe or a voltage-temperature probe. In both cases, TRS is broken by the magnetic flux. However, the voltage-temperature probe requires an additional anisotropy in the system for its TRS-breaking effects to significantly influence transport. We examine the role of this anisotropy in enhancing performance. Our results show that the EMP and efficiency at a given power can be enhanced by increasing the strength of nonlinear power dissipation, even though the output power remains unchanged. The voltage probe configuration generally yields higher power, while the voltage-temperature probe is more efficient, except in certain regimes where large asymmetries and high figures of merit allow the voltage probe setup to outperform. | [
{
"version": "v1",
"created": "Wed, 23 Oct 2024 16:55:49 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 03:17:37 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Behera",
"Jayasmita",
""
],
[
"Bedkihal",
"Salil",
""
],
[
"Agarwalla",
"Bijay Kumar",
""
],
[
"Bandyopadhyay",
"Malay",
""
]
]
|
2410.19075 | Volodymyr Maistrenko | Oleksandr Sudakov and Volodymyr Maistrenko | Parallelization of Network Dynamics Computations in Heterogeneous Distributed Environment | 15 pages, 14 figures | null | null | null | nlin.CD cs.DC nlin.PS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of parallelizing computations to study non-linear dynamics in large networks of non-locally coupled oscillators using heterogeneous computing resources. The proposed approach can be applied to a variety of non-linear dynamics models with runtime specification of parameters and network topologies. Parallelizing the solution of equations for different network elements is performed transparently and, in contrast to available tools, does not require parallel programming from end-users. The runtime scheduler takes into account the performance of computing and communication resources to reduce downtime and to achieve a quasi-optimal parallelizing speed-up. The proposed approach was implemented, and its efficiency is proven by numerous applications for simulating large dynamical networks with 10^3-10^8 elements described by Hodgkin-Huxley, FitzHugh-Nagumo, and Kuramoto models, for investigating pathological synchronization during Parkinson's disease, analyzing multi-stability, for studying chimera and solitary states in 3D networks, etc. All the above computations may be performed using symmetrical multiprocessors, graphic processing units, and a network of workstations within the same run and it was demonstrated that near-linear speed-up can be achieved for large networks. The proposed approach is promising for extension to new hardware like edge-computing devices. | [
{
"version": "v1",
"created": "Thu, 24 Oct 2024 18:32:04 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 18:41:43 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Sudakov",
"Oleksandr",
""
],
[
"Maistrenko",
"Volodymyr",
""
]
]
|
2410.20160 | Gabriele Dessena | Gabriele Dessena, Marco Civera, Andr\'es Marcos, Bernardino Chiaia and Oscar E. Bonilla-Manrique | Vibration-based damage detection of a trainer jet via multiple input tangential interpolation | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | Vibration-based damage detection relies on identifying modal parameters, such as natural frequencies and mode shapes, to be used as structural integrity indicators. However, traditional comparisons of these parameters are often ambiguous in complex systems, complicating damage detection and assessment. The modified total modal assurance criterion (MTMAC), a metric well-known in the field of finite element model updating, is extended to address this challenge and is proposed as a metric for damage identification and severity assessment. To support the requirement for precise and robust modal identification of Structural Health Monitoring (SHM), the improved Loewner Framework (iLF), known for its reliability and computational performance, is pioneeringly employed within SHM. Since the MTMAC is proposed solely as a damage identification and severity assessment metric, the coordinate modal assurance criterion (COMAC), also a well-established tool, but for damage localisation using mode shapes, is used for completeness. The iLF SHM capabilities are validated through comparisons with traditional methods, including least-squares complex exponential (LSCE) and stochastic subspace identification with canonical variate analysis (SSI-CVA) on a numerical case study of a cantilever beam. Furthermore, the MTMAC is validated against the traditional vibration-based approach, which involves directly comparing natural frequencies and mode shapes. Finally, an experimental dataset from a BAE Systems Hawk T1A trainer jet ground vibration tests is used to demonstrate the iLF and MTMAC capabilities on real-life, real-size SHM problem, showing their effectiveness in detecting and assessing damage. | [
{
"version": "v1",
"created": "Sat, 26 Oct 2024 12:22:14 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 10:49:12 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Dessena",
"Gabriele",
""
],
[
"Civera",
"Marco",
""
],
[
"Marcos",
"Andrés",
""
],
[
"Chiaia",
"Bernardino",
""
],
[
"Bonilla-Manrique",
"Oscar E.",
""
]
]
|
2410.21015 | Pankaj Mishra | Nalinikanta Pradhan, Rina Kanamoto, M. Bhattacharya, and Pankaj Kumar Mishra | Signature of Andreev-Bashkin superfluid drag from Cavity Optomechanics | Results with lattice rotation have been added. Now the cavity sepectrum reveals more pronounced Andreev-Bashkin superfluid drag | Physical Review Research 7, 023051 (2025) | 10.1103/PhysRevResearch.7.023051 | null | cond-mat.quant-gas | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Andreev-Bashkin (AB) effect, corresponding to the dissipationless dragging of one superfluid by another, was predicted almost fifty years ago but has so far eluded experimental detection. In this work, we theoretically introduce an entirely new detection paradigm to this quest, and show that it enables the observation of the hitherto undetected AB effect for realistic parameters. We accomplish this by using the powerful techniques of cavity optomechanics, which were crucial to the observation of gravitational waves, on a spinor ring Bose-Einstein condensate. In contrast to all known AB detection methods, our scheme allows for real-time, \textit{in situ}, minimally destructive and three orders-of-magnitude more sensitive measurement of the AB effect. Our proposal, which considers persistent currents in weakly repulsive atomic condensates, and amplifies the AB signal using a novel dynamic Bragg spectroscopy technique, is supported by numerical simulations of the stochastic Gross-Pitaevski equation, which agree very well with our analytic Bogoliubov-de Gennes calculations. Our work suggests a novel tool for sensitively and nondestructively probing the dynamics of rotationally interacting superfluids using cavities and has fundamental implications for ongoing studies of superfluid hydrodynamics, atomtronics, matter-wave interferometry, and cavity optomechanical sensing. | [
{
"version": "v1",
"created": "Mon, 28 Oct 2024 13:38:19 GMT"
},
{
"version": "v2",
"created": "Wed, 04 Dec 2024 07:27:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Pradhan",
"Nalinikanta",
""
],
[
"Kanamoto",
"Rina",
""
],
[
"Bhattacharya",
"M.",
""
],
[
"Mishra",
"Pankaj Kumar",
""
]
]
|
2410.21553 | Shaorong Zhang | Shaorong Zhang, Yuanbin Cheng, Greg Ver Steeg | Exploring the Design Space of Diffusion Bridge Models | 23 pages, 9 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Diffusion bridge models and stochastic interpolants enable high-quality image-to-image (I2I) translation by creating paths between distributions in pixel space. However, the proliferation of techniques based on incompatible mathematical assumptions have impeded progress. In this work, we unify and expand the space of bridge models by extending Stochastic Interpolants (SIs) with preconditioning, endpoint conditioning, and an optimized sampling algorithm. These enhancements expand the design space of diffusion bridge models, leading to state-of-the-art performance in both image quality and sampling efficiency across diverse I2I tasks. Furthermore, we identify and address a previously overlooked issue of low sample diversity under fixed conditions. We introduce a quantitative analysis for output diversity and demonstrate how we can modify the base distribution for further improvements. | [
{
"version": "v1",
"created": "Mon, 28 Oct 2024 21:30:59 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 18:24:40 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Zhang",
"Shaorong",
""
],
[
"Cheng",
"Yuanbin",
""
],
[
"Steeg",
"Greg Ver",
""
]
]
|
2410.21878 | Qing Ping Liu | Lingling Xue, Shasha Wang and Qing Ping Liu | B\"acklund-Darboux Transformations for Super KdV Type Equations | null | SIGMA 21 (2025), 050, 22 pages | 10.3842/SIGMA.2025.050 | null | nlin.SI | http://creativecommons.org/licenses/by-sa/4.0/ | By introducing a Miura transformation, we derive a generalized super modified Korteweg-de Vries (gsmKdV) equation from the generalized super KdV (gsKdV) equation. It is demonstrated that, while the gsKdV equation takes Kupershmidt's super KdV (sKdV) equation and Geng-Wu's sKdV equation as two distinct reductions, there are also two equations, namely Kupershmidt's super modified KdV (smKdV) equation and Hu's smKdV equation, which are associated with the gsmKdV equation. By analyzing the flows within the gsKdV and gsmKdV hierarchies, we specifically derive the first negative flows associated with both hierarchies.We then construct a number of B\"acklund-Darboux transformations (BDTs) for both the gsKdV and gsmKdV equations, elucidating the interrelationship between them. By proper reductions, we are able not only to recover the previously known BDTs for Kupershimdt's sKdV and smKdV equations, but also to obtain the BDTs for the Geng-Wu's sKdV/smKdV and Hu's smKdV equations. As applications, we construct some exact solutions for those equations. Since all flows of the sKdV or smKdV hierarchy share the same spatial parts of spectral problem, thus these Darboux matrices and spatial parts of BTs are applicable to any flow of those hierarchies. | [
{
"version": "v1",
"created": "Tue, 29 Oct 2024 09:13:37 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 05:31:08 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Xue",
"Lingling",
""
],
[
"Wang",
"Shasha",
""
],
[
"Liu",
"Qing Ping",
""
]
]
|
2410.23202 | Jiaying Yang | Jiaying Yang, Maryam Khanahmadi, Ingrid Strandberg, Akshay Gaikwad, Claudia Castillo-Moreno, Anton Frisk Kockum, Muhammad Asad Ullah, G\"oran Johansson, Axel Martin Eriksson, and Simone Gasparinetti | Deterministic generation of frequency-bin-encoded microwave photons | null | Physical Review Letters 134, 240803 (2025) | 10.1103/PhysRevLett.134.240803 | null | quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | A distributed quantum computing network requires a quantum communication channel between spatially separated processing units. In superconducting circuits, such a channel can be implemented based on propagating microwave photons to encode and transfer quantum information between an emitter and a receiver. However, traveling microwave photons can be lost during the transmission, leading to the failure of information transfer. Heralding protocols can be used to detect such photon losses. In this work, we propose such a protocol and experimentally demonstrate a frequency-bin encoding method of microwave photonic modes using superconducting circuits. We deterministically encode the quantum information from a superconducting qubit by simultaneously emitting its information into two photonic modes at different frequencies, with a process fidelity of 90.4%. The frequency-bin-encoded photonic modes can be used, at the receiver processor, to detect the occurrence of photon loss. Our work thus provides a reliable method to implement high-fidelity quantum state transfer in a distributed quantum computing network, incorporating error detection to enhance performance and accuracy. | [
{
"version": "v1",
"created": "Wed, 30 Oct 2024 16:53:01 GMT"
},
{
"version": "v2",
"created": "Mon, 23 Jun 2025 15:44:56 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yang",
"Jiaying",
""
],
[
"Khanahmadi",
"Maryam",
""
],
[
"Strandberg",
"Ingrid",
""
],
[
"Gaikwad",
"Akshay",
""
],
[
"Castillo-Moreno",
"Claudia",
""
],
[
"Kockum",
"Anton Frisk",
""
],
[
"Ullah",
"Muhammad Asad",
""
],
[
"Johansson",
"Göran",
""
],
[
"Eriksson",
"Axel Martin",
""
],
[
"Gasparinetti",
"Simone",
""
]
]
|
2411.00603 | Anna Nickolaevna Morozovska | Oleksandr S. Pylypchuk, Victor V. Vainberg, Vladimir N. Poroshin, Anastasyia V. Terebilenko, Andrii S. Nikolenko, Vadym I. Popenko, Anatolyi S. Tolochko, Maryna V. Olenchuk, Oleksii Bezkrovnyi, Galina I. Dovbeshko, Tomash Sabov, Boris M. Romanyuk, Sergey V. Kolotilov, and Anna N. Morozovska | The impact of morphological structure and flexo-chemical strains on the electric transport mechanisms in the molybdenum-disulfide-oxide nanoflakes | 39 pages, including 9 figures and 3 tables, 60 references and Supplementary Materials.Submitted to the Journal of Applied Physics | null | null | null | cond-mat.mtrl-sci cond-mat.mes-hall | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Electric conduction mechanisms are studied in the pressed nanoflake powder of the molybdenum-disulfide-oxide (MoSxOy) depending on their content and structure. The MoSxOy nanoflakes were prepared by reaction of (NH4)6Mo7O24 with thiourea in aqueous solution followed by aerial oxidation. The sintered nanoflakes are 10-20 nm thick and self-assembled in the "nanoflower"-shape aggregates forming powder particles. The chemical composition and structure of the powders were studied by XPS, EDS and Raman spectroscopy, which show that the powders have different chemical composition and structure depending on the preparation conditions. These studies revealed the existence of different forms of MoS2 and its oxides in the powders. These features are impactful on electric transport properties. The current-voltage (I-V) curves of the pressed MoSxOy nanoflakes reveal hysteresis-like behavior; at that the loop width depends on the chemical composition and structure. The negative differential conductivity is observed for the highest content of Mo in oxide/sulfoxide form. The I-V curves of all MoSxOy nanoflake samples manifest the three-state resistive switching and the long-lasting transient charge/discharge on switching "on/off" the voltage across the sample, which evidences the role of interface charges in their conductivity. To describe theoretically the observed I-V curves, polar and electric-transport properties of the pressed MoSxOy nanoflakes, the Landau-Cahn-Hilliard approach considering flexo-chemical field has been used. The revealed experimentally and explained theoretically features of resistive switching and charge accumulation look promising for applications in memristors and high-performance supercapacitors. | [
{
"version": "v1",
"created": "Fri, 01 Nov 2024 14:07:05 GMT"
},
{
"version": "v2",
"created": "Sat, 16 Nov 2024 13:21:33 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Apr 2025 21:05:04 GMT"
},
{
"version": "v4",
"created": "Thu, 03 Jul 2025 13:21:46 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Pylypchuk",
"Oleksandr S.",
""
],
[
"Vainberg",
"Victor V.",
""
],
[
"Poroshin",
"Vladimir N.",
""
],
[
"Terebilenko",
"Anastasyia V.",
""
],
[
"Nikolenko",
"Andrii S.",
""
],
[
"Popenko",
"Vadym I.",
""
],
[
"Tolochko",
"Anatolyi S.",
""
],
[
"Olenchuk",
"Maryna V.",
""
],
[
"Bezkrovnyi",
"Oleksii",
""
],
[
"Dovbeshko",
"Galina I.",
""
],
[
"Sabov",
"Tomash",
""
],
[
"Romanyuk",
"Boris M.",
""
],
[
"Kolotilov",
"Sergey V.",
""
],
[
"Morozovska",
"Anna N.",
""
]
]
|
2411.00863 | Chenyang An | Chenyang An, Shima Imani, Feng Yao, Chengyu Dong, Ali Abbasi, Harsh Shrivastava, Samuel Buss, Jingbo Shang, Gayathri Mahalingam, Pramod Sharma, Maurice Diesendruck | Next-Token Prediction Task Assumes Optimal Data Ordering for LLM Training in Proof Generation | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the field of large language model (LLM)-based proof generation, despite extensive training on large datasets such as ArXiv, LLMs still exhibit only modest performance on proving tasks of moderate difficulty. We believe that this is partly due to the widespread presence of suboptimal ordering within the data for each proof used in training. For example, published proofs often follow a purely logical order, where each step logically proceeds from the previous steps based on the deductive rules. This order is designed to facilitate the verification of the proof's soundness, rather than to help people and models learn the discovery process of the proof. In proof generation, we argue that the optimal order for one training data sample occurs when the relevant intermediate supervision for a particular proof step in the proof is always positioned to the left of that proof step. We call such order the intuitively sequential order. We validate our claims using two tasks: intuitionistic propositional logic theorem-proving and digit multiplication. Our experiments verify the order effect and provide support for our explanations. We demonstrate that training is most effective when the proof is in the intuitively sequential order. Moreover, the order effect and the performance gap between models trained on different data orders can be substantial -- with an 11 percent improvement in proof success rate observed in the propositional logic theorem-proving task, between models trained on the optimal order compared to the worst order. Lastly, we define a common type of order issue in advanced math proofs and find that 17.3 percent of theorems with nontrivial proofs in the first two chapters of a widely used graduate-level mathematics textbook suffer from this issue. A detailed list of those proofs is provided in the appendix. | [
{
"version": "v1",
"created": "Wed, 30 Oct 2024 18:00:04 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 15:14:51 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"An",
"Chenyang",
""
],
[
"Imani",
"Shima",
""
],
[
"Yao",
"Feng",
""
],
[
"Dong",
"Chengyu",
""
],
[
"Abbasi",
"Ali",
""
],
[
"Shrivastava",
"Harsh",
""
],
[
"Buss",
"Samuel",
""
],
[
"Shang",
"Jingbo",
""
],
[
"Mahalingam",
"Gayathri",
""
],
[
"Sharma",
"Pramod",
""
],
[
"Diesendruck",
"Maurice",
""
]
]
|
2411.00909 | Mustafa Halilsoy | M. Halilsoy and V. Memari | Extremal Black Holes in Cosmology from Colliding Light Beams | 10 pages | null | null | null | gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | From the duality between the interaction region of colliding waves and black holes (BHs) it is known that certain BH solutions can be obtained from colliding gravity coupled electromagnetic (em) waves. In the limit of vanishing gravity waves, we show that extremal Reissner-Nordstrom (RN) BH can form. We show also that direct collision of pure em waves creates the near horizon geometry (NHG) of the same BH, supporting our point. Due to the quantum restrictions such BHs can have horizon radii in the range $R>10^{8}m$. If such BHs from pure light do exist this may open a new frontier in our understanding of cosmological BHs at large. The generation of impulsive gravitational waves in the process may be instrumental in the detection of a BH created from pure light. | [
{
"version": "v1",
"created": "Fri, 01 Nov 2024 11:18:51 GMT"
},
{
"version": "v2",
"created": "Tue, 19 Nov 2024 10:44:13 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 06:46:56 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Halilsoy",
"M.",
""
],
[
"Memari",
"V.",
""
]
]
|
2411.02136 | Robert Fonod | Robert Fonod and Haechan Cho and Hwasoo Yeo and Nikolas Geroliminis | Advanced computer vision for extracting georeferenced vehicle trajectories from drone imagery | null | Transportation Research Part C: Emerging Technologies, Vol. 178, p. 105205, 2025 | 10.1016/j.trc.2025.105205 | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | This paper presents a framework for extracting georeferenced vehicle trajectories from high-altitude drone imagery, addressing key challenges in urban traffic monitoring and the limitations of traditional ground-based systems. Our approach integrates several novel contributions, including a tailored object detector optimized for high-altitude bird's-eye view perspectives, a unique track stabilization method that uses detected vehicle bounding boxes as exclusion masks during image registration, and an orthophoto and master frame-based georeferencing strategy that enhances consistent alignment across multiple drone viewpoints. Additionally, our framework features robust vehicle dimension estimation and detailed road segmentation, enabling comprehensive traffic analysis. Conducted in the Songdo International Business District, South Korea, the study utilized a multi-drone experiment covering 20 intersections, capturing approximately 12TB of 4K video data over four days. The framework produced two high-quality datasets: the Songdo Traffic dataset, comprising approximately 700,000 unique vehicle trajectories, and the Songdo Vision dataset, containing over 5,000 human-annotated images with about 300,000 vehicle instances in four classes. Comparisons with high-precision sensor data from an instrumented probe vehicle highlight the accuracy and consistency of our extraction pipeline in dense urban environments. The public release of Songdo Traffic and Songdo Vision, and the complete source code for the extraction pipeline, establishes new benchmarks in data quality, reproducibility, and scalability in traffic research. Results demonstrate the potential of integrating drone technology with advanced computer vision for precise and cost-effective urban traffic monitoring, providing valuable resources for developing intelligent transportation systems and enhancing traffic management strategies. | [
{
"version": "v1",
"created": "Mon, 04 Nov 2024 14:49:01 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 09:25:50 GMT"
},
{
"version": "v3",
"created": "Wed, 25 Jun 2025 20:45:19 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Fonod",
"Robert",
""
],
[
"Cho",
"Haechan",
""
],
[
"Yeo",
"Hwasoo",
""
],
[
"Geroliminis",
"Nikolas",
""
]
]
|
2411.02697 | Minho Choi | Minho Choi, Jinlin Xiang, Anna Wirth-Singh, Seung-Hwan Baek, Eli Shlizerman, and Arka Majumdar | Transferable polychromatic optical encoder for neural networks | 21 pages, 4 figures, 2 tables | Nat Commun 16, 5623 (2025) | 10.1038/s41467-025-61338-4 | null | cs.CV physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Artificial neural networks (ANNs) have fundamentally transformed the field of computer vision, providing unprecedented performance. However, these ANNs for image processing demand substantial computational resources, often hindering real-time operation. In this paper, we demonstrate an optical encoder that can perform convolution simultaneously in three color channels during the image capture, effectively implementing several initial convolutional layers of a ANN. Such an optical encoding results in ~24,000 times reduction in computational operations, with a state-of-the art classification accuracy (~73.2%) in free-space optical system. In addition, our analog optical encoder, trained for CIFAR-10 data, can be transferred to the ImageNet subset, High-10, without any modifications, and still exhibits moderate accuracy. Our results evidence the potential of hybrid optical/digital computer vision system in which the optical frontend can pre-process an ambient scene to reduce the energy and latency of the whole computer vision system. | [
{
"version": "v1",
"created": "Tue, 05 Nov 2024 00:49:47 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Choi",
"Minho",
""
],
[
"Xiang",
"Jinlin",
""
],
[
"Wirth-Singh",
"Anna",
""
],
[
"Baek",
"Seung-Hwan",
""
],
[
"Shlizerman",
"Eli",
""
],
[
"Majumdar",
"Arka",
""
]
]
|
2411.02928 | Quinten Eggerickx | Quinten Eggerickx and Adam Wills and Ting-Chun Lin and Kristiaan De Greve and Min-Hsiu Hsieh | Almost Linear Decoder for Optimal Geometrically Local Quantum Codes | 19 pages, 9 figures. We corrected the noise model from circuit-level to code-capacity. An incorrect threshold proof was removed; we now prove a threshold for MWPM and argue applicability to our decoder. We added motivation and discuss applicability to codes in arXiv:2303.06755 and arXiv:2309.16503. Typos have been fixed | Physical Review Research, 7(2), 023300 (2025) | 10.1103/sxdx-qbcz | null | quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Geometrically local quantum codes, which are error correction codes embedded in $\mathbb{R}^D$ with checks acting only on qubits within a fixed spatial distance, have garnered significant interest. Recently, it has been demonstrated how to achieve geometrically local codes that maximize both the dimension and the distance, as well as the energy barrier of the code. In this work, we focus on the constructions involving subdivision and show that they have an almost linear time decoder, obtained by combining the decoder of the outer good qLDPC code and a generalized version of the Union-Find decoder. This provides the first decoder for an optimal geometrically local three-dimensional code. We demonstrate the existence of a finite threshold error rate under the code capacity noise model using a minimum weight perfect matching decoder. Furthermore, we argue that this threshold is also applicable to the decoder based on the generalized Union-Find algorithm. | [
{
"version": "v1",
"created": "Tue, 05 Nov 2024 09:15:06 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 15:40:00 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Eggerickx",
"Quinten",
""
],
[
"Wills",
"Adam",
""
],
[
"Lin",
"Ting-Chun",
""
],
[
"De Greve",
"Kristiaan",
""
],
[
"Hsieh",
"Min-Hsiu",
""
]
]
|
2411.03208 | Clement de Chaisemartin | Facundo Arga\~naraz, Cl\'ement de Chaisemartin, Ziteng Lei | Randomly Assigned First Differences? | null | null | null | null | econ.EM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider treatment-effect estimation using a first-difference regression of an outcome evolution $\Delta Y$ on a treatment evolution $\Delta D$. Under a causal model in levels with a time-varying effect, the regression residual is a function of the period-one treatment $D_{1}$. Then, researchers should test if $\Delta D$ and $D_{1}$ are correlated: if they are, the regression may suffer from an omitted variable bias. To solve it, researchers may control nonparametrically for $E(\Delta D|D_{1})$. We use our results to revisit first-difference regressions estimated on the data of \cite{acemoglu2016import}, who study the effect of imports from China on US employment. $\Delta D$ and $D_{1}$ are strongly correlated, thus implying that first-difference regressions may be biased if the effect of Chinese imports changes over time. The coefficient on $\Delta D$ is no longer significant when controlling for $E(\Delta D|D_{1})$. | [
{
"version": "v1",
"created": "Tue, 05 Nov 2024 15:58:39 GMT"
},
{
"version": "v2",
"created": "Fri, 08 Nov 2024 16:36:00 GMT"
},
{
"version": "v3",
"created": "Mon, 02 Dec 2024 00:08:59 GMT"
},
{
"version": "v4",
"created": "Fri, 20 Dec 2024 10:25:36 GMT"
},
{
"version": "v5",
"created": "Thu, 09 Jan 2025 09:57:45 GMT"
},
{
"version": "v6",
"created": "Wed, 29 Jan 2025 15:48:39 GMT"
},
{
"version": "v7",
"created": "Thu, 03 Jul 2025 14:33:06 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Argañaraz",
"Facundo",
""
],
[
"de Chaisemartin",
"Clément",
""
],
[
"Lei",
"Ziteng",
""
]
]
|
2411.03419 | Henning Schl\"omer | Henning Schl\"omer, Annabelle Bohrdt, Fabian Grusdt | Geometric orthogonal metals: Hidden antiferromagnetism and pseudogap from fluctuating stripes | 14 pages, 4 figures | null | null | null | cond-mat.str-el cond-mat.quant-gas | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | One of the key features of hole-doped cuprates is the presence of an extended pseudogap phase, whose microscopic origin has been the subject of intense investigation since its discovery and is believed to be crucial for understanding high-temperature superconductivity. Various explanations have been proposed for the pseudogap, including links to symmetry-breaking orders such as stripes or pairing, and the emergence of novel fractionalized Fermi liquid (FL*) and orthogonal metal (OM) phases. The topological nature of FL* and OM phases has been identified as scenarios compatible with a small Fermi surface without symmetry breaking, as suggested experimentally. With recent experimental and numerical studies supporting an intricate relationship between stripe order and the pseudogap phase, we here propose an alternative scenario: an orthogonal metal with a geometric origin (GOM) driven by fluctuating domain walls. The essential mechanism behind our proposal is hidden order, where the proliferation of domain walls stabilized by charge fluctuations obscures the underlying long-range antiferromagnetic order in real-space, but order is preserved in the reference frame of the background spins. As a result, well-defined fermionic quasiparticles in the form of magnetic polarons exist, which couple to $\mathbb{Z}_2$ topological excitations of the domain wall string-net condensate in the ground state and constitute a small Fermi surface. At a critical doping value, we argue that hidden order is lost, driving a transition to a regular Fermi liquid at a hidden quantum critical point (hQCP) featuring quantum critical transport properties. Our GOM framework establishes a deep connection between the antiferromagnetic, stripe, and pseudogap phases, and suggests a possible unification of superconductivity in (electron and hole) doped cuprates and heavy fermion compounds. | [
{
"version": "v1",
"created": "Tue, 05 Nov 2024 18:59:15 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 10:28:42 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Schlömer",
"Henning",
""
],
[
"Bohrdt",
"Annabelle",
""
],
[
"Grusdt",
"Fabian",
""
]
]
|
2411.03748 | Gleb Fedorovich | Gleb Fedorovich, Clemens Kuhlenkamp, Atac Imamoglu and Ivan Amelio | First order Quantum Hall to Wigner crystal phase transition on a triangular lattice: an iDMRG study | null | null | null | null | cond-mat.str-el cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | In this work we study a system of interacting fermions on a triangular lattice in the presence of an external magnetic field. We neglect spin and fix a density of one third, with one unit of magnetic flux per particle. The infinite density matrix renormalization group algorithm is used to compute the ground state of this generalized Fermi-Hubbard model. Increasing the strength of the nearest-neighbor repulsion, we find a first order transition between an Integer Quantum Hall phase and a crystalline, generalized Wigner crystal state. The first-order nature of the phase transition is consistent with a Ginzburg-Landau argument. We expect our results to be relevant for moir\'e heterostructures of two-dimensional materials. | [
{
"version": "v1",
"created": "Wed, 06 Nov 2024 08:24:32 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 14:32:02 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Fedorovich",
"Gleb",
""
],
[
"Kuhlenkamp",
"Clemens",
""
],
[
"Imamoglu",
"Atac",
""
],
[
"Amelio",
"Ivan",
""
]
]
|
2411.05197 | Cheng Zhang | Cheng Zhang, Hanna Foerster, Robert D. Mullins, Yiren Zhao, Ilia Shumailov | Hardware and Software Platform Inference | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | It is now a common business practice to buy access to large language model (LLM) inference rather than self-host, because of significant upfront hardware infrastructure and energy costs. However, as a buyer, there is no mechanism to verify the authenticity of the advertised service including the serving hardware platform, e.g. that it is actually being served using an NVIDIA H100. Furthermore, there are reports suggesting that model providers may deliver models that differ slightly from the advertised ones, often to make them run on less expensive hardware. That way, a client pays premium for a capable model access on more expensive hardware, yet ends up being served by a (potentially less capable) cheaper model on cheaper hardware. In this paper we introduce hardware and software platform inference (HSPI) -- a method for identifying the underlying GPU architecture and software stack of a (black-box) machine learning model solely based on its input-output behavior. Our method leverages the inherent differences of various GPU architectures and compilers to distinguish between different GPU types and software stacks. By analyzing the numerical patterns in the model's outputs, we propose a classification framework capable of accurately identifying the GPU used for model inference as well as the underlying software configuration. Our findings demonstrate the feasibility of inferring GPU type from black-box models. We evaluate HSPI against models served on different real hardware and find that in a white-box setting we can distinguish between different GPUs with between $83.9\%$ and $100\%$ accuracy. Even in a black-box setting we achieve results that are up to 3x higher than random guess accuracy. Our code is available at https://github.com/ChengZhang-98/HSPI. | [
{
"version": "v1",
"created": "Thu, 07 Nov 2024 21:44:04 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 22:14:13 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Zhang",
"Cheng",
""
],
[
"Foerster",
"Hanna",
""
],
[
"Mullins",
"Robert D.",
""
],
[
"Zhao",
"Yiren",
""
],
[
"Shumailov",
"Ilia",
""
]
]
|
2411.05670 | Sebasti\'an Carrasco | Sebastian C. Carrasco, Sean Lourette, Ignacio Sola, and Vladimir S. Malinovsky | Dynamically Enhanced Two-Photon Spectroscopy | null | Phys. Rev. Lett. 134, 163601 (2025) | 10.1103/PhysRevLett.134.163601 | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A novel quantum control protocol utilizing two-photon processes with trigonometric pulse modulation is developed, enabling the intermediate state population's dynamic elimination (DE). The proposed DE technique excels at single-photon resonance in contrast to the well-known adiabatic elimination (AE) regime, which requires large single-photon detuning and strong pulses. The DE approach outperforms AE in efficiency and exhibits enhanced resilience to one-photon detuning since the key control parameter, the effective two-photon Rabi frequency, inversely proportional to the modulation frequency, does not depend on the one-photon detuning. The comprehensive analysis of population transfer and Ramsey interferometry demonstrates the protocol's superiority, achieving enhanced signal-to-noise ratios and higher fidelities with respect to existing two-photon methods. | [
{
"version": "v1",
"created": "Fri, 08 Nov 2024 16:16:26 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Carrasco",
"Sebastian C.",
""
],
[
"Lourette",
"Sean",
""
],
[
"Sola",
"Ignacio",
""
],
[
"Malinovsky",
"Vladimir S.",
""
]
]
|
2411.06285 | Peiran Xiao | Peiran Xiao | Allocating Positional Goods: A Mechanism Design Approach | null | null | null | null | econ.TH | http://creativecommons.org/licenses/by/4.0/ | I study the optimal allocation of positional goods in the presence of externalities arising from consumers' concerns for relative consumption. Using a mechanism design approach, I characterize the externalities by a feasibility condition. I show that the revenue-maximizing mechanism possibly excludes some buyers and fully separates the participants if their type distribution satisfies Myerson's regularity. The seller can guarantee at least half the maximum revenue by offering a single level of positional goods. When there is no exclusion, as the seller offers more positional good levels, the consumer surplus decreases (increases) if the distribution has an increasing (decreasing) failure rate. Higher participation increases the consumer surplus under increasing failure rates. Applications include luxury goods, priority services, education, and organizational hierarchies. | [
{
"version": "v1",
"created": "Sat, 09 Nov 2024 21:02:57 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 06:36:37 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Xiao",
"Peiran",
""
]
]
|
2411.06854 | Qi Song | Qi Song, Jie Lou, Yan Chen | Condensation of Composite Bosonic Trions in Interacting Bose-Fermi Mixtures | 6 pages, 7 figures | null | null | null | cond-mat.quant-gas | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We reveal a quantum coherent state characterized by composite bosonic trions, wherein paired fermions further bind with bosons, in one-dimensional Bose-Fermi mixtures.This phase emerges in two separate models, both featuring onsite boson-fermion attraction that induces negative binding energy for the composite trions. The first is the pair-hopping model, in which spinless fermions undergo pair hopping to form preformed pairs. The second is the extended Bose-Fermi Hubbard model, describing dipolar bosons and fermions with density-density interactions. Notably, the formation of composite trions is independent of the specific pairing interactions of fermions. Extensive density matrix renormalization group calculations demonstrate the quasi-condensation of the composite trions, evidenced by algebraically decaying correlations of trions, gapped single-particle excitations, and suppressed fermion pair correlations. Our findings provide valuable insights into the three-body pairing mechanism in mixed-particle systems. | [
{
"version": "v1",
"created": "Mon, 11 Nov 2024 10:37:02 GMT"
},
{
"version": "v2",
"created": "Sun, 08 Dec 2024 09:02:46 GMT"
},
{
"version": "v3",
"created": "Sat, 12 Apr 2025 06:04:11 GMT"
},
{
"version": "v4",
"created": "Thu, 03 Jul 2025 14:08:04 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Song",
"Qi",
""
],
[
"Lou",
"Jie",
""
],
[
"Chen",
"Yan",
""
]
]
|
2411.07618 | Linyi Yang | Qingyu Yin, Chak Tou Leong, Hongbo Zhang, Minjun Zhu, Hanqi Yan, Qiang Zhang, Yulan He, Wenjie Li, Jun Wang, Yue Zhang, Linyi Yang | Direct Preference Optimization Using Sparse Feature-Level Constraints | null | ICML 2025 | null | null | cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | The alignment of large language models (LLMs) with human preferences remains a key challenge. While post-training techniques like Reinforcement Learning from Human Feedback (RLHF) and Direct Preference Optimization (DPO) have achieved notable success, they often introduce computational inefficiencies and training instability. In this paper, we propose Feature-level constrained Preference Optimization (FPO), a novel method designed to simplify the alignment process while ensuring stability. FPO leverages pre-trained Sparse Autoencoders (SAEs) and introduces feature-level constraints, allowing for efficient, sparsity-enforced alignment. Our approach enjoys efficiency by using sparse features activated in a well-trained sparse autoencoder and the quality of sequential KL divergence by using the feature-level offline reference. Experimental results on benchmark datasets demonstrate that FPO achieves a 5.08% absolute improvement in win rate with much lower computational cost compared to state-of-the-art baselines, making it a promising solution for efficient and controllable LLM alignments. | [
{
"version": "v1",
"created": "Tue, 12 Nov 2024 07:54:13 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 13:18:23 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yin",
"Qingyu",
""
],
[
"Leong",
"Chak Tou",
""
],
[
"Zhang",
"Hongbo",
""
],
[
"Zhu",
"Minjun",
""
],
[
"Yan",
"Hanqi",
""
],
[
"Zhang",
"Qiang",
""
],
[
"He",
"Yulan",
""
],
[
"Li",
"Wenjie",
""
],
[
"Wang",
"Jun",
""
],
[
"Zhang",
"Yue",
""
],
[
"Yang",
"Linyi",
""
]
]
|
2411.08777 | Pit Henrich | Pit Henrich, Franziska Mathis-Ullrich, Paul Maria Scheikl | LUDO: Low-Latency Understanding of Deformable Objects using Point Cloud Occupancy Functions | Published in IEEE Transactions on Robotics (T-RO) | null | 10.1109/TRO.2025.3582837 | null | cs.RO cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately determining the shape of deformable objects and the location of their internal structures is crucial for medical tasks that require precise targeting, such as robotic biopsies. We introduce LUDO, a method for accurate low-latency understanding of deformable objects. LUDO reconstructs objects in their deformed state, including their internal structures, from a single-view point cloud observation in under 30 ms using occupancy networks. LUDO provides uncertainty estimates for its predictions. Additionally, it provides explainability by highlighting key features in its input observations. Both uncertainty and explainability are important for safety-critical applications such as surgery. We evaluate LUDO in real-world robotic experiments, achieving a success rate of 98.9% for puncturing various regions of interest (ROIs) inside deformable objects. We compare LUDO to a popular baseline and show its superior ROI localization accuracy, training time, and memory requirements. LUDO demonstrates the potential to interact with deformable objects without the need for deformable registration methods. | [
{
"version": "v1",
"created": "Wed, 13 Nov 2024 17:02:46 GMT"
},
{
"version": "v2",
"created": "Wed, 09 Apr 2025 17:47:52 GMT"
},
{
"version": "v3",
"created": "Sat, 19 Apr 2025 12:33:15 GMT"
},
{
"version": "v4",
"created": "Thu, 08 May 2025 07:18:17 GMT"
},
{
"version": "v5",
"created": "Thu, 03 Jul 2025 15:01:00 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Henrich",
"Pit",
""
],
[
"Mathis-Ullrich",
"Franziska",
""
],
[
"Scheikl",
"Paul Maria",
""
]
]
|
2411.09851 | Ho Fung Tsoi | Ho Fung Tsoi, Dylan Rankin, Cecile Caillol, Miles Cranmer, Sridhara Dasu, Javier Duarte, Philip Harris, Elliot Lipeles, Vladimir Loncar | SymbolFit: Automatic Parametric Modeling with Symbolic Regression | 52 pages, 35 figures. Under review. The API can be used out-of-the-box and is available at https://github.com/hftsoi/symbolfit | Comput. Softw. Big Sci. 9, 12 (2025) | 10.1007/s41781-025-00140-9 | null | hep-ex cs.LG physics.data-an | http://creativecommons.org/licenses/by/4.0/ | We introduce SymbolFit, a framework that automates parametric modeling by using symbolic regression to perform a machine-search for functions that fit the data while simultaneously providing uncertainty estimates in a single run. Traditionally, constructing a parametric model to accurately describe binned data has been a manual and iterative process, requiring an adequate functional form to be determined before the fit can be performed. The main challenge arises when the appropriate functional forms cannot be derived from first principles, especially when there is no underlying true closed-form function for the distribution. In this work, we develop a framework that automates and streamlines the process by utilizing symbolic regression, a machine learning technique that explores a vast space of candidate functions without requiring a predefined functional form because the functional form itself is treated as a trainable parameter, making the process far more efficient and effortless than traditional regression methods. We demonstrate the framework in high-energy physics experiments at the CERN Large Hadron Collider (LHC) using five real proton-proton collision datasets from new physics searches, including background modeling in resonance searches for high-mass dijet, trijet, paired-dijet, diphoton, and dimuon events. We show that our framework can flexibly and efficiently generate a wide range of candidate functions that fit a nontrivial distribution well using a simple fit configuration that varies only by random seed, and that the same fit configuration, which defines a vast function space, can also be applied to distributions of different shapes, whereas achieving a comparable result with traditional methods would have required extensive manual effort. | [
{
"version": "v1",
"created": "Fri, 15 Nov 2024 00:09:37 GMT"
},
{
"version": "v2",
"created": "Tue, 11 Feb 2025 02:11:22 GMT"
},
{
"version": "v3",
"created": "Sun, 02 Mar 2025 23:29:50 GMT"
},
{
"version": "v4",
"created": "Sun, 11 May 2025 02:19:58 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Tsoi",
"Ho Fung",
""
],
[
"Rankin",
"Dylan",
""
],
[
"Caillol",
"Cecile",
""
],
[
"Cranmer",
"Miles",
""
],
[
"Dasu",
"Sridhara",
""
],
[
"Duarte",
"Javier",
""
],
[
"Harris",
"Philip",
""
],
[
"Lipeles",
"Elliot",
""
],
[
"Loncar",
"Vladimir",
""
]
]
|
2411.10860 | Santiago Guzm\'an-Pro | Manuel Bodirsky and Santiago Guzm\'an-Pro | Hereditary First-Order Logic: the tractable quantifier prefix classes | The second version focuses on hereditary first-order logic, and for a better streamlined presentation, we removed all content on extensional ESO which will be submitted elsewhere. Version 2 also contains new results (Section 4) | null | null | null | math.LO cs.CC cs.DM cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many computational problems can be modelled as the class of all finite structures $\mathbb A$ that satisfy a fixed first-order sentence $\phi$ hereditarily, i.e., we require that every (induced) substructure of $\mathbb A$ satisfies $\phi$. We call the corresponding computational problem the hereditary model checking problem for $\phi$, and denote it by Her$(\phi)$.
We present a complete description of the quantifier prefixes for $\phi$ such that Her$(\phi)$ is in P; we show that for every other quantifier prefix there exists a formula $\phi$ with this prefix such that Her$(\phi)$ is coNP-complete. Specifically, we show that if $Q$ is of the form $\forall^\ast\exists\forall^\ast$ or of the form $\forall^\ast\exists^\ast$, then Her$(\phi)$ can be solved in polynomial time whenever the quantifier prefix of $\phi$ is $Q$. Otherwise, $Q$ contains $\exists \exists \forall$ or $\exists \forall \exists$ as a subword, and in this case, there is a first-order formula $\phi$ whose quantifier prefix is $Q$ and Her$(\phi)$ is coNP-complete. Moreover, we show that there is no algorithm that decides for a given first-order formula $\phi$ whether Her$(\phi)$ is in P (unless P$=$NP). | [
{
"version": "v1",
"created": "Sat, 16 Nov 2024 18:50:15 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 08:14:16 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Bodirsky",
"Manuel",
""
],
[
"Guzmán-Pro",
"Santiago",
""
]
]
|
2411.11042 | Javier Rivera-Dean | J. Rivera-Dean, P. Stammer, M. F. Ciappina and M. Lewenstein | Structured squeezed light allows for high-harmonic generation in classical forbidden geometries | 30 pages (8 main text + 22 Supplementary Material), 16 figures (5 main text + 11 Supplementary Material). Comments are welcome. The latest version corresponds to the final revised version of the manuscript | Physical Review Letters 135, 013801 (2025) | 10.1103/4hdl-bdwj | null | quant-ph physics.atom-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | High-harmonic generation (HHG) is a nonlinear process in which a strong driving field interacts with a material, resulting in the frequency up-conversion of the driver into its high-order harmonics. This process is highly sensitive to the field's polarization: circular polarization, for instance, inhibits HHG. In this work, we demonstrate that the use of non-classical structured light enables HHG in this otherwise prohibitive configuration for classical drivers. We consider circularly polarized light with non-classical fluctuations, introduced via squeezing along one polarization direction, and show that these non-classical features prompt the HHG process. We find that the spectral properties of the emitted harmonics depend on the type of squeezing applied and, by analyzing the inner electron dynamics, we relate the observed differences to modifications of the HHG three-step mechanism induced by the specific squeezing type. This approach opens new pathways for integrating quantum optics in HHG, providing novel means of controlling the light-matter interaction dynamics. | [
{
"version": "v1",
"created": "Sun, 17 Nov 2024 11:19:44 GMT"
},
{
"version": "v2",
"created": "Mon, 06 Jan 2025 20:34:21 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 07:32:07 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Rivera-Dean",
"J.",
""
],
[
"Stammer",
"P.",
""
],
[
"Ciappina",
"M. F.",
""
],
[
"Lewenstein",
"M.",
""
]
]
|
2411.11382 | Mudassir Ibrahim Awan | Mudassir Ibrahim Awan, Ahsan Raza, Waseem Hassan, Ki-Uk Kyung, and Seokhee Jeon | Quantifying Haptic Affection of Car Door through Data-Driven Analysis of Force Profile | 12 pages, 9 figures, 3 tables. Mudassir Ibrahim Awan and Ahsan Raza are equally contributing authors | null | 10.1109/ACCESS.2025.3585067 | null | cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Haptic affection plays a crucial role in user experience, particularly in the automotive industry where the tactile quality of components can influence customer satisfaction. This study aims to accurately predict the affective property of a car door by only watching the force or torque profile of it when opening. To this end, a deep learning model is designed to capture the underlying relationships between force profiles and user-defined adjective ratings, providing insights into the door-opening experience. The dataset employed in this research includes force profiles and user adjective ratings collected from six distinct car models, reflecting a diverse set of door-opening characteristics and tactile feedback. The model's performance is assessed using Leave-One-Out Cross-Validation, a method that measures its generalization capability on unseen data. The results demonstrate that the proposed model achieves a high level of prediction accuracy, indicating its potential in various applications related to haptic affection and design optimization in the automotive industry. | [
{
"version": "v1",
"created": "Mon, 18 Nov 2024 08:58:53 GMT"
},
{
"version": "v2",
"created": "Fri, 02 May 2025 05:49:56 GMT"
},
{
"version": "v3",
"created": "Thu, 22 May 2025 08:11:09 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Awan",
"Mudassir Ibrahim",
""
],
[
"Raza",
"Ahsan",
""
],
[
"Hassan",
"Waseem",
""
],
[
"Kyung",
"Ki-Uk",
""
],
[
"Jeon",
"Seokhee",
""
]
]
|
2411.13638 | Martino Stefanini | Yi-Fan Qu, Martino Stefanini, Tao Shi, Tilman Esslinger, Sarang Gopalakrishnan, Jamir Marino and Eugene Demler | Variational approach to the dynamics of dissipative quantum impurity models | 19 pages, 7 figures. Close to published version. Editors' Suggestion | Phys. Rev. B 111, 155113 (2025) | 10.1103/PhysRevB.111.155113 | null | cond-mat.quant-gas cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent experiments with quantum simulators using ultracold atoms and superconducting qubits have demonstrated the potential of controlled dissipation as a versatile tool for realizing correlated many-body states. However, determining the dynamics of dissipative quantum many-body systems remains a significant analytical and numerical challenge. In this work, we focus on a dissipative impurity problem as a testbed for new methodological developments. We introduce an efficient non-perturbative framework that combines the superposition of Gaussian states (SGS) variational ansatz with the quantum trajectory approach to simulate open systems featuring a dissipative impurity. Applying this method to a spinful impurity subject to two-body losses and embedded in a bath of noninteracting fermions, we explore the full crossover from weak to strong dissipation regimes. The non-perturbative nature of the SGS ansatz allows us to thoroughly examine this crossover, providing comprehensive insights into the system's behavior. In the strong dissipation regime, our approach reproduces the finding that localized two-body losses can induce the Kondo effect [arXiv:2406.03527], characterized by a slowdown of spin relaxation and an enhancement of charge conductance. Furthermore, we reveal an exotic ``negative conductance" phenomenon at zero potential bias -- a counter-intuitive single-body effect resulting from intermediate dissipation and finite bandwidth. Finally, we investigate the formation of ferromagnetic domains and propose an extension to realize a higher-spin Kondo model using localized dissipation. | [
{
"version": "v1",
"created": "Wed, 20 Nov 2024 19:00:00 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 13:07:10 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Qu",
"Yi-Fan",
""
],
[
"Stefanini",
"Martino",
""
],
[
"Shi",
"Tao",
""
],
[
"Esslinger",
"Tilman",
""
],
[
"Gopalakrishnan",
"Sarang",
""
],
[
"Marino",
"Jamir",
""
],
[
"Demler",
"Eugene",
""
]
]
|
2411.14441 | Kahraman Kostas Dr | Kahraman Kostas, Rabia Yasa Kostas, Mike Just, Michael A. Lones | GeMID: Generalizable Models for IoT Device Identification | 11 pages main, 24 pages Supplementary Material, 35 pages total | null | null | null | cs.CR cs.AI cs.NI | http://creativecommons.org/licenses/by/4.0/ | With the proliferation of devices on the Internet of Things (IoT), ensuring their security has become paramount. Device identification (DI), which distinguishes IoT devices based on their traffic patterns, plays a crucial role in both differentiating devices and identifying vulnerable ones, closing a serious security gap. However, existing approaches to DI that build machine learning models often overlook the challenge of model generalizability across diverse network environments. In this study, we propose a novel framework to address this limitation and to evaluate the generalizability of DI models across data sets collected within different network environments. Our approach involves a two-step process: first, we develop a feature and model selection method that is more robust to generalization issues by using a genetic algorithm with external feedback and datasets from distinct environments to refine the selections. Second, the resulting DI models are then tested on further independent datasets to robustly assess their generalizability. We demonstrate the effectiveness of our method by empirically comparing it to alternatives, highlighting how fundamental limitations of commonly employed techniques such as sliding window and flow statistics limit their generalizability. Moreover, we show that statistical methods, widely used in the literature, are unreliable for device identification due to their dependence on network-specific characteristics rather than device-intrinsic properties, challenging the validity of a significant portion of existing research. Our findings advance research in IoT security and device identification, offering insight into improving model effectiveness and mitigating risks in IoT networks. | [
{
"version": "v1",
"created": "Tue, 05 Nov 2024 17:09:43 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 21:26:35 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Kostas",
"Kahraman",
""
],
[
"Kostas",
"Rabia Yasa",
""
],
[
"Just",
"Mike",
""
],
[
"Lones",
"Michael A.",
""
]
]
|
2411.14604 | Salvatore Federico | Salvatore Federico, Fausto Gozzi, Andrzej \'Swi\k{e}ch | On Mean Field Games in Infinite Dimension | null | null | null | null | math.AP math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study a Mean Field Games (MFG) system in a real, separable infinite dimensional Hilbert space. The system consists of a second order parabolic type equation, called Hamilton-Jacobi-Bellman (HJB) equation in the paper, coupled with a nonlinear Fokker-Planck (FP) equation. Both equations contain a Kolmogorov operator. Solutions to the HJB equation are interpreted in the mild solution sense and solutions to the FP equation are interpreted in an appropriate weak sense. We prove well-posedness of the considered MFG system under certain conditions. The existence of a solution to the MFG system is proved using Tikhonov's fixed point theorem in a proper space. Uniqueness of solutions is obtained under typical separability and Lasry-Lions type monotonicity conditions. | [
{
"version": "v1",
"created": "Thu, 21 Nov 2024 22:08:22 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 09:58:37 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Federico",
"Salvatore",
""
],
[
"Gozzi",
"Fausto",
""
],
[
"Święch",
"Andrzej",
""
]
]
|
2411.16100 | Christoph Treude | Christoph Treude and Christopher M. Poskitt | Bot-Driven Development: From Simple Automation to Autonomous Software Development Bots | BotSE 2025: International Workshop on Bots in Software Engineering | Proc. BotSE'25, pages 18-22. IEEE, 2025 | 10.1109/BotSE67031.2025.00012 | null | cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As software development increasingly adopts automation, bot-driven development (BotDD) represents a transformative shift where bots assume proactive roles in coding, testing, and project management. In bot-driven development, bots go beyond support tasks, actively driving development workflows by making autonomous decisions, performing independent assessments, and managing code quality and dependencies. This paper explores how bot-driven development impacts traditional development roles, particularly in redefining driver-navigator dynamics, and aligns with DevOps goals for faster feedback, continuous learning, and efficiency. We propose a research agenda addressing challenges in bot-driven development, including skill development for developers, human-bot trust dynamics, optimal interruption frequency, and ethical considerations. Through empirical studies and prototype systems, our aim is to define best practices and governance structures for integrating bot-driven development into modern software engineering. | [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 05:21:23 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Treude",
"Christoph",
""
],
[
"Poskitt",
"Christopher M.",
""
]
]
|
2411.16765 | Shester Gueuwou | Shester Gueuwou, Xiaodan Du, Greg Shakhnarovich, Karen Livescu, Alexander H. Liu | SHuBERT: Self-Supervised Sign Language Representation Learning via Multi-Stream Cluster Prediction | Fixed Figure 1. ACL 2025 | null | null | null | cs.CL cs.CV | http://creativecommons.org/licenses/by/4.0/ | Sign language processing has traditionally relied on task-specific models, limiting the potential for transfer learning across tasks. Pre-training methods for sign language have typically focused on either supervised pre-training, which cannot take advantage of unlabeled data, or context-independent (frame or video segment) representations, which ignore the effects of relationships across time in sign language. We introduce SHuBERT (Sign Hidden-Unit BERT), a self-supervised contextual representation model learned from approximately 1,000 hours of American Sign Language video. SHuBERT adapts masked token prediction objectives to multi-stream visual sign language input, learning to predict multiple targets corresponding to clustered hand, face, and body pose streams. SHuBERT achieves state-of-the-art performance across multiple tasks including sign language translation, isolated sign language recognition, and fingerspelling detection. | [
{
"version": "v1",
"created": "Mon, 25 Nov 2024 03:13:08 GMT"
},
{
"version": "v2",
"created": "Tue, 03 Jun 2025 01:30:30 GMT"
},
{
"version": "v3",
"created": "Wed, 02 Jul 2025 23:43:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Gueuwou",
"Shester",
""
],
[
"Du",
"Xiaodan",
""
],
[
"Shakhnarovich",
"Greg",
""
],
[
"Livescu",
"Karen",
""
],
[
"Liu",
"Alexander H.",
""
]
]
|
2411.17623 | Edmond Chaussidon | E. Chaussidon, C. Y\`eche, A. de Mattia, C. Payerne, P. McDonald, A. J. Ross, S. Ahlen, D. Bianchi, D. Brooks, E. Burtin, T. Claybaugh, A. de la Macorra, P. Doel, S. Ferraro, A. Font-Ribera, J. E. Forero-Romero, E. Gazta\~naga, H. Gil-Mar\'in, S. Gontcho A Gontcho, G. Gutierrez, J. Guy, K. Honscheid, C. Howlett, D. Huterer, R. Kehoe, D. Kirkby, T. Kisner, A. Kremin, L. Le Guillou, M. E. Levi, M. Manera, A. Meisner, R. Miquel, J. Moustakas, J. A. Newman, G. Niz, N. Palanque-Delabrouille, W. J. Percival, F. Prada, I. P\'erez-R\`afols, C. Ravoux, G. Rossi, E. Sanchez, D. Schlegel, M. Schubnell, H. Seo, D. Sprayberry, G. Tarl\'e, M. Vargas-Maga\~na, B. A. Weaver, C. Zhao, H. Zou | Constraining primordial non-Gaussianity with DESI 2024 LRG and QSO samples | null | null | null | null | astro-ph.CO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We analyse the large-scale clustering of the Luminous Red Galaxy (LRG) and Quasar (QSO) sample from the first data release (DR1) of the Dark Energy Spectroscopic Instrument (DESI). In particular, we constrain the primordial non-Gaussianity (PNG) parameter $f_{\rm NL}^{\rm loc}$ via the large-scale scale-dependent bias in the power spectrum using $1,631,716$ LRGs ($0.6 < z < 1.1$) and $1,189,129$ QSOs ($0.8 < z < 3.1$). This new measurement takes advantage of the enormous statistical power at large scales of DESI DR1 data, surpassing the latest data release (DR16) of the extended Baryon Oscillation Spectroscopic Survey (eBOSS). For the first time in this kind of analysis, we use a blinding procedure to mitigate the risk of confirmation bias in our results. We improve the model of the radial integral constraint proposing an innovative correction of the window function. We also carefully test the mitigation of the dependence of the target selection on the photometry qualities by incorporating an angular integral constraint contribution to the window function, and validate our methodology with the blinded data. Finally, combining the two samples, we measure $f_{\rm NL}^{\rm loc} = {-3.6}_{-9.1}^{+9.0}$ at $68\%$ confidence, where we assume the universality relation for the LRG sample and a recent merger model for the QSO sample about the response of bias to primordial non-Gaussianity. Adopting the universality relation for the PNG bias in the QSO analysis leads to $f_{\rm NL}^{\rm loc} = 3.5_{-7.4}^{+10.7}$ at $68\%$ confidence. This measurement is the most precise determination of primordial non-Gaussianity using large-scale structure to date, surpassing the latest result from eBOSS by a factor of $2.3$. | [
{
"version": "v1",
"created": "Tue, 26 Nov 2024 17:34:01 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 21:49:19 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Chaussidon",
"E.",
""
],
[
"Yèche",
"C.",
""
],
[
"de Mattia",
"A.",
""
],
[
"Payerne",
"C.",
""
],
[
"McDonald",
"P.",
""
],
[
"Ross",
"A. J.",
""
],
[
"Ahlen",
"S.",
""
],
[
"Bianchi",
"D.",
""
],
[
"Brooks",
"D.",
""
],
[
"Burtin",
"E.",
""
],
[
"Claybaugh",
"T.",
""
],
[
"de la Macorra",
"A.",
""
],
[
"Doel",
"P.",
""
],
[
"Ferraro",
"S.",
""
],
[
"Font-Ribera",
"A.",
""
],
[
"Forero-Romero",
"J. E.",
""
],
[
"Gaztañaga",
"E.",
""
],
[
"Gil-Marín",
"H.",
""
],
[
"Gontcho",
"S. Gontcho A",
""
],
[
"Gutierrez",
"G.",
""
],
[
"Guy",
"J.",
""
],
[
"Honscheid",
"K.",
""
],
[
"Howlett",
"C.",
""
],
[
"Huterer",
"D.",
""
],
[
"Kehoe",
"R.",
""
],
[
"Kirkby",
"D.",
""
],
[
"Kisner",
"T.",
""
],
[
"Kremin",
"A.",
""
],
[
"Guillou",
"L. Le",
""
],
[
"Levi",
"M. E.",
""
],
[
"Manera",
"M.",
""
],
[
"Meisner",
"A.",
""
],
[
"Miquel",
"R.",
""
],
[
"Moustakas",
"J.",
""
],
[
"Newman",
"J. A.",
""
],
[
"Niz",
"G.",
""
],
[
"Palanque-Delabrouille",
"N.",
""
],
[
"Percival",
"W. J.",
""
],
[
"Prada",
"F.",
""
],
[
"Pérez-Ràfols",
"I.",
""
],
[
"Ravoux",
"C.",
""
],
[
"Rossi",
"G.",
""
],
[
"Sanchez",
"E.",
""
],
[
"Schlegel",
"D.",
""
],
[
"Schubnell",
"M.",
""
],
[
"Seo",
"H.",
""
],
[
"Sprayberry",
"D.",
""
],
[
"Tarlé",
"G.",
""
],
[
"Vargas-Magaña",
"M.",
""
],
[
"Weaver",
"B. A.",
""
],
[
"Zhao",
"C.",
""
],
[
"Zou",
"H.",
""
]
]
|
2411.18339 | Esfandiar Nava-Yazdani | Esfandiar Nava-Yazdani | Ridge Regression for Manifold-valued Time-Series with Application to Meteorological Forecast | null | null | null | null | math.DG cs.NA math.NA stat.AP stat.ML | http://creativecommons.org/licenses/by/4.0/ | We propose a natural intrinsic extension of the ridge regression from Euclidean spaces to general manifolds, which relies on Riemannian least-squares fitting, empirical covariance, and Mahalanobis distance. We utilize it for time-series prediction and apply the approach to forecast hurricane tracks and their wind speeds. | [
{
"version": "v1",
"created": "Wed, 27 Nov 2024 13:36:43 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 15:03:23 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Nava-Yazdani",
"Esfandiar",
""
]
]
|
2411.18849 | Paul Egr\'e | Paul \'Egr\'e and Ellie Ripley | Probabilistic consequence relations | Published in Journal of Logic and Computation | null | 10.1093/logcom/exae076 | null | math.LO cs.LO math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates logical consequence defined in terms of probability distributions, for a classical propositional language using a standard notion of probability. We examine three distinct probabilistic consequence notions, which we call material consequence, preservation consequence, and symmetric consequence. While material consequence is fully classical for any threshold, preservation consequence and symmetric consequence are subclassical, with only symmetric consequence gradually approaching classical logic at the limit threshold equal to 1. Our results extend earlier results obtained by J. Paris in a SET-FMLA setting to the SET-SET setting, and consider open thresholds beside closed ones. In the SET-SET setting, in particular, they reveal that probability 1 preservation does not yield classical logic, but supervaluationism, and conversely positive probability preservation yields subvaluationism. | [
{
"version": "v1",
"created": "Thu, 28 Nov 2024 01:32:22 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 02:18:42 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Égré",
"Paul",
""
],
[
"Ripley",
"Ellie",
""
]
]
|
2411.19688 | Kim-Celine Kahl | Kim-Celine Kahl, Selen Erkan, Jeremias Traub, Carsten T. L\"uth, Klaus Maier-Hein, Lena Maier-Hein, Paul F. Jaeger | SURE-VQA: Systematic Understanding of Robustness Evaluation in Medical VQA Tasks | TMLR 07/2025 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision-Language Models (VLMs) have great potential in medical tasks, like Visual Question Answering (VQA), where they could act as interactive assistants for both patients and clinicians. Yet their robustness to distribution shifts on unseen data remains a key concern for safe deployment. Evaluating such robustness requires a controlled experimental setup that allows for systematic insights into the model's behavior. However, we demonstrate that current setups fail to offer sufficiently thorough evaluations. To address this gap, we introduce a novel framework, called SURE-VQA, centered around three key requirements to overcome current pitfalls and systematically analyze VLM robustness: 1) Since robustness on synthetic shifts does not necessarily translate to real-world shifts, it should be measured on real-world shifts that are inherent to the VQA data; 2) Traditional token-matching metrics often fail to capture underlying semantics, necessitating the use of large language models (LLMs) for more accurate semantic evaluation; 3) Model performance often lacks interpretability due to missing sanity baselines, thus meaningful baselines should be reported that allow assessing the multimodal impact on the VLM. To demonstrate the relevance of this framework, we conduct a study on the robustness of various Fine-Tuning (FT) methods across three medical datasets with four types of distribution shifts. Our study highlights key insights into robustness: 1) No FT method consistently outperforms others in robustness, and 2) robustness trends are more stable across FT methods than across distribution shifts. Additionally, we find that simple sanity baselines that do not use the image data can perform surprisingly well and confirm LoRA as the best-performing FT method on in-distribution data. Code is provided at https://github.com/IML-DKFZ/sure-vqa. | [
{
"version": "v1",
"created": "Fri, 29 Nov 2024 13:22:52 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 14:30:53 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 13:07:30 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Kahl",
"Kim-Celine",
""
],
[
"Erkan",
"Selen",
""
],
[
"Traub",
"Jeremias",
""
],
[
"Lüth",
"Carsten T.",
""
],
[
"Maier-Hein",
"Klaus",
""
],
[
"Maier-Hein",
"Lena",
""
],
[
"Jaeger",
"Paul F.",
""
]
]
|
2412.00420 | Lan Feng | Lan Feng, Fan Nie, Yuejiang Liu, Alexandre Alahi | TAROT: Targeted Data Selection via Optimal Transport | null | null | null | null | cs.LG cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose TAROT, a targeted data selection framework grounded in optimal transport theory. Previous targeted data selection methods primarily rely on influence-based greedy heuristics to enhance domain-specific performance. While effective on limited, unimodal data (i.e., data following a single pattern), these methods struggle as target data complexity increases. Specifically, in multimodal distributions, these heuristics fail to account for multiple inherent patterns, leading to suboptimal data selection. This work identifies two primary factors contributing to this limitation: (i) the disproportionate impact of dominant feature components in high-dimensional influence estimation, and (ii) the restrictive linear additive assumptions inherent in greedy selection strategies. To address these challenges, TAROT incorporates whitened feature distance to mitigate dominant feature bias, providing a more reliable measure of data influence. Building on this, TAROT uses whitened feature distance to quantify and minimize the optimal transport distance between the selected data and target domains. Notably, this minimization also facilitates the estimation of optimal selection ratios. We evaluate TAROT across multiple tasks, including semantic segmentation, motion prediction, and instruction tuning. Results consistently show that TAROT outperforms state-of-the-art methods, highlighting its versatility across various deep learning tasks. Code is available at https://github.com/vita-epfl/TAROT. | [
{
"version": "v1",
"created": "Sat, 30 Nov 2024 10:19:51 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 19:48:29 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Feng",
"Lan",
""
],
[
"Nie",
"Fan",
""
],
[
"Liu",
"Yuejiang",
""
],
[
"Alahi",
"Alexandre",
""
]
]
|
2412.01940 | Vihan Lakshman | Blaise Munyampirwa, Vihan Lakshman, Benjamin Coleman | Down with the Hierarchy: The 'H' in HNSW Stands for "Hubs" | 17 pages | null | null | null | cs.LG cs.DB cs.IR | http://creativecommons.org/licenses/by/4.0/ | Driven by recent breakthrough advances in neural representation learning, approximate near-neighbor (ANN) search over vector embeddings has emerged as a critical computational workload. With the introduction of the seminal Hierarchical Navigable Small World (HNSW) algorithm, graph-based indexes have established themselves as the overwhelmingly dominant paradigm for efficient and scalable ANN search. As the name suggests, HNSW searches a layered hierarchical graph to quickly identify neighborhoods of similar points to a given query vector. But is this hierarchy even necessary? A rigorous experimental analysis to answer this question would provide valuable insights into the nature of algorithm design for ANN search and motivate directions for future work in this increasingly crucial domain. We conduct an extensive benchmarking study covering more large-scale datasets than prior investigations of this question. We ultimately find that a flat navigable small world graph graph retains all of the benefits of HNSW on high-dimensional datasets, with latency and recall performance essentially \emph{identical} to the original algorithm but with less memory overhead. Furthermore, we go a step further and study \emph{why} the hierarchy of HNSW provides no benefit in high dimensions, hypothesizing that navigable small world graphs contain a well-connected, frequently traversed ``highway" of hub nodes that maintain the same purported function as the hierarchical layers. We present compelling empirical evidence that the \emph{Hub Highway Hypothesis} holds for real datasets and investigate the mechanisms by which the highway forms. The implications of this hypothesis may also provide future research directions in developing enhancements to graph-based ANN search. | [
{
"version": "v1",
"created": "Mon, 02 Dec 2024 20:04:06 GMT"
},
{
"version": "v2",
"created": "Sun, 02 Feb 2025 19:25:13 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 13:32:02 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Munyampirwa",
"Blaise",
""
],
[
"Lakshman",
"Vihan",
""
],
[
"Coleman",
"Benjamin",
""
]
]
|
2412.03349 | Alexandre Fournier Montgieux | Alexandre Fournier-Montgieux, Michael Soumm, Adrian Popescu, Bertrand Luvison, Herv\'e Le Borgne | Fairer Analysis and Demographically Balanced Face Generation for Fairer Face Verification | Published in WACV2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Face recognition and verification are two computer vision tasks whose performances have advanced with the introduction of deep representations. However, ethical, legal, and technical challenges due to the sensitive nature of face data and biases in real-world training datasets hinder their development. Generative AI addresses privacy by creating fictitious identities, but fairness problems remain. Using the existing DCFace SOTA framework, we introduce a new controlled generation pipeline that improves fairness. Through classical fairness metrics and a proposed in-depth statistical analysis based on logit models and ANOVA, we show that our generation pipeline improves fairness more than other bias mitigation approaches while slightly improving raw performance. | [
{
"version": "v1",
"created": "Wed, 04 Dec 2024 14:30:19 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 12:42:44 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Fournier-Montgieux",
"Alexandre",
""
],
[
"Soumm",
"Michael",
""
],
[
"Popescu",
"Adrian",
""
],
[
"Luvison",
"Bertrand",
""
],
[
"Borgne",
"Hervé Le",
""
]
]
|
2412.04165 | Shobhna Misra | Shobhna Misra and Reshma Peremadathil Pradeep and Yaoxuan Feng and Urs Grob and Andrada Oana Mandru and Christian L. Degen and Hans J. Hug and Alexander Eichler | Differential Magnetic Force Microscopy with a Switchable Tip | 7 pages, 4 figures in Main and 3 pages, 3 figures in supplement | null | 10.1103/yky5-qcrk | null | physics.app-ph cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | The separation of physical forces acting on the tip of a magnetic force microscope (MFM) is essential for correct magnetic imaging. Electrostatic forces can be modulated by varying the tip-sample potential and minimized to map the local Kelvin potential. However, distinguishing magnetic forces from van der Waals forces typically requires two measurements with opposite tip magnetizations under otherwise identical measurement conditions. Here, we present an inverted magnetic force microscope where the sample is mounted on a flat cantilever for force sensing, and the magnetic tip is attached to a miniaturized electromagnet that periodically flips the tip magnetization. This setup enables the extraction of magnetic tip-sample interactions from the sidebands occurring at the switching rate in the cantilever oscillation spectrum. Our method achieves the separation of magnetic signals from other force contributions in a single-scan mode. Future iterations of this setup may incorporate membrane, trampoline, or string resonators with ultra-high quality factors, potentially improving measurement sensitivity by up to three orders of magnitude compared to the state-of-the-art MFM systems using cantilevers. | [
{
"version": "v1",
"created": "Thu, 05 Dec 2024 14:03:15 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Misra",
"Shobhna",
""
],
[
"Pradeep",
"Reshma Peremadathil",
""
],
[
"Feng",
"Yaoxuan",
""
],
[
"Grob",
"Urs",
""
],
[
"Mandru",
"Andrada Oana",
""
],
[
"Degen",
"Christian L.",
""
],
[
"Hug",
"Hans J.",
""
],
[
"Eichler",
"Alexander",
""
]
]
|
2412.04795 | Masaaki Harada | Masaaki Harada | Ternary near-extremal self-dual codes of lengths $36$, $48$ and $60$ | 13 pages | null | null | null | cs.IT math.CO math.IT | http://creativecommons.org/licenses/by/4.0/ | For lengths $36$, $48$ and $60$, we construct new ternary near-extremal self-dual codes with weight enumerators for which no ternary near-extremal self-dual codes were previously known to exist. | [
{
"version": "v1",
"created": "Fri, 06 Dec 2024 06:06:59 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 07:24:10 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Harada",
"Masaaki",
""
]
]
|
2412.05693 | Michael Metel R | Michael R. Metel, Boxing Chen, Mehdi Rezagholizadeh | Batch-Max: Higher LLM Throughput using Larger Batch Sizes and KV Cache Compression | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Several works have developed eviction policies to remove key-value (KV) pairs from the KV cache for more efficient inference. The focus has been on compressing the KV cache after the input prompt has been processed for faster token generation. In settings with limited GPU memory, and when the input context is longer than the generation length, we show that by also compressing the KV cache during the input processing phase, larger batch sizes can be used resulting in significantly higher throughput while still maintaining the original model's accuracy. | [
{
"version": "v1",
"created": "Sat, 07 Dec 2024 16:41:54 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Jun 2025 02:24:51 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 16:06:35 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Metel",
"Michael R.",
""
],
[
"Chen",
"Boxing",
""
],
[
"Rezagholizadeh",
"Mehdi",
""
]
]
|
2412.05827 | Tiancheng Li | Tiancheng Li, Weijian Luo, Zhiyang Chen, Liyuan Ma, Guo-Jun Qi | Self-Guidance: Boosting Flow and Diffusion Generation on Their Own | 16 pages, 10 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Proper guidance strategies are essential to achieve high-quality generation results without retraining diffusion and flow-based text-to-image models. Existing guidance either requires specific training or strong inductive biases of diffusion model networks, potentially limiting their applications. Motivated by the observation that artifact outliers can be detected by a significant decline in the density from a noisier to a cleaner noise level, we propose Self-Guidance (SG), which improves the image quality by suppressing the generation of low-quality samples. SG only relies on the sampling probabilities of its own diffusion model at different noise levels with no need of any guidance-specific training. This makes it flexible to be used in a plug-and-play manner with other sampling algorithms. We also introduce a more efficient approximation of SG, named SG-prev, which reuses the output from the immediately previous diffusion step to avoid doubling sampling time. We conduct experiments on text-to-image and text-to-video generation with different architectures, including UNet and transformer models. With open-sourced diffusion models such as Stable Diffusion 3.5 and FLUX, SG exceeds existing algorithms on multiple metrics, including both FID and Human Preference Score. SG-prev also achieves strong results over both the baseline and the SG with only one forward pass. Moreover, we find that SG and SG-prev both have a surprisingly positive effect on the generation of physiologically correct human body structures such as hands, faces, and arms, showing their ability of eliminating human body artifacts with minimal efforts. We will release our code along with this paper. | [
{
"version": "v1",
"created": "Sun, 08 Dec 2024 06:32:27 GMT"
},
{
"version": "v2",
"created": "Sat, 08 Mar 2025 13:10:47 GMT"
},
{
"version": "v3",
"created": "Sun, 25 May 2025 05:42:43 GMT"
},
{
"version": "v4",
"created": "Thu, 03 Jul 2025 04:33:09 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Li",
"Tiancheng",
""
],
[
"Luo",
"Weijian",
""
],
[
"Chen",
"Zhiyang",
""
],
[
"Ma",
"Liyuan",
""
],
[
"Qi",
"Guo-Jun",
""
]
]
|
2412.06889 | Alberto Salvio | Indra Kumar Banerjee, Francesco Rescigno, Alberto Salvio | Primordial Black Holes (as Dark Matter) from the Supercooled Phase Transitions with Radiative Symmetry Breaking | v2: 30 pages, 10 figures, more explanations related to the primordial black hole production mechanism added, version accepted for publication in JCAP | null | null | null | hep-ph astro-ph.CO gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | We study in detail the production of primordial black holes (PBHs), as well as their mass and initial spin, due to the phase transitions corresponding to radiative symmetry breaking (RSB) and featuring a large supercooling. The latter property allows us to use a model-independent approach. In this context, we demonstrate that the decay rate of the false vacuum grows exponentially with time to a high degree of accuracy, justifying a time dependence commonly assumed in the literature. Our study provides ready-to-use results for determining the abundance, mass and initial spin of PBHs generated in a generic RSB model with large supercooling. We find that PBHs are generically produced in a broad region of the model-independent parameter space. As an application, we identify the subregion that may explain recently reported microlensing anomalies. Additionally, we show that a simple Standard-Model extension, with right-handed neutrinos and gauged $B-L$ featuring RSB, may explain an anomaly of this sort in a region of its parameter space. | [
{
"version": "v1",
"created": "Mon, 09 Dec 2024 19:00:02 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 09:50:22 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Banerjee",
"Indra Kumar",
""
],
[
"Rescigno",
"Francesco",
""
],
[
"Salvio",
"Alberto",
""
]
]
|
2412.06946 | Osvaldo Gramaxo Freitas | Osvaldo Gramaxo Freitas, Anastasios Theodoropoulos, Nino Villanueva, Tiago Fernandes, Solange Nunes, Jos\'e A. Font, Antonio Onofre, Alejandro Torres-Forn\'e, Jos\'e D. Martin-Guerrero | A Deep Learning Powered Numerical Relativity Surrogate for Binary Black Hole Waveforms | null | null | null | null | gr-qc astro-ph.HE astro-ph.IM cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gravitational-wave approximants are essential for gravitational-wave astronomy, allowing the coverage binary black hole parameter space for inference or match filtering without costly numerical relativity (NR) simulations, but generally trading some accuracy for computational efficiency. To reduce this trade-off, NR surrogate models can be constructed using interpolation within NR waveform space. We present a 2-stage training approach for neural network-based NR surrogate models. Initially trained on approximant-generated waveforms and then fine-tuned with NR data, these dual-stage artificial neural surrogate (\texttt{DANSur}) models offer rapid and competitively accurate waveform generation, generating millions in under 20ms on a GPU while keeping mean mismatches with NR around $10^{-4}$. Implemented in the \textsc{bilby} framework, we show they can be used for parameter estimation tasks. | [
{
"version": "v1",
"created": "Mon, 09 Dec 2024 19:45:30 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 08:56:15 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Freitas",
"Osvaldo Gramaxo",
""
],
[
"Theodoropoulos",
"Anastasios",
""
],
[
"Villanueva",
"Nino",
""
],
[
"Fernandes",
"Tiago",
""
],
[
"Nunes",
"Solange",
""
],
[
"Font",
"José A.",
""
],
[
"Onofre",
"Antonio",
""
],
[
"Torres-Forné",
"Alejandro",
""
],
[
"Martin-Guerrero",
"José D.",
""
]
]
|
2412.07056 | Matthias Franz | Matthias Franz | Fibrations of classifying spaces in the simplicial setting | 5 pages | null | 10.1515/gmj-2025-2013 | null | math.AT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this note we show that in the simplicial setting, the classifying space construction converts short exact sequences of groups not just to homotopy fibrations, but in fact to fibre bundles. | [
{
"version": "v1",
"created": "Mon, 09 Dec 2024 23:37:16 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Franz",
"Matthias",
""
]
]
|
2412.07844 | Edoardo Ballini | Edoardo Ballini, Julius Mildenberger, Matteo M. Wauters, Philipp Hauke | Symmetry verification for noisy quantum simulations of non-Abelian lattice gauge theories | 24 pages, 8 figures | null | null | null | quant-ph cond-mat.other hep-lat nucl-th | http://creativecommons.org/licenses/by/4.0/ | Non-Abelian gauge theories underlie our understanding of fundamental forces of modern physics. Simulating them on quantum hardware is an outstanding challenge in the rapidly evolving field of quantum simulation. A key prerequisite is the protection of local gauge symmetries against errors that, if unchecked, would lead to unphysical results. While an extensive toolkit devoted to identifying, mitigating, and ultimately correcting such errors has been developed for Abelian groups, non-commuting symmetry operators complicate the implementation of similar schemes in non-Abelian theories. Here, we discuss two techniques for error mitigation through symmetry verification, tailored for non-Abelian lattice gauge theories implemented in noisy qudit hardware: dynamical post-selection (DPS), based on mid-circuit measurements without active feedback, and post-processed symmetry verification (PSV), which combines measurements of correlations between target observables and gauge transformations. We illustrate both approaches for the discrete non-Abelian group $D_3$ in 2+1 dimensions, explaining their usefulness for current NISQ devices even in the presence of fast fluctuating noise. Our results open new avenues for robust quantum simulation of non-Abelian gauge theories, for further development of error-mitigation techniques, and for measurement-based control methods in qudit platforms. | [
{
"version": "v1",
"created": "Tue, 10 Dec 2024 19:00:02 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 14:09:02 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Ballini",
"Edoardo",
""
],
[
"Mildenberger",
"Julius",
""
],
[
"Wauters",
"Matteo M.",
""
],
[
"Hauke",
"Philipp",
""
]
]
|
2412.08589 | Pranava Seth | Pranava Seth, Vishal Upendran, Megha Anand, Janmejoy Sarkar, Soumya Roy, Priyadarshan Chaki, Pratyay Chowdhury, Borishan Ghosh, Durgesh Tripathi | SPACE-SUIT: An Artificial Intelligence Based Chromospheric Feature Extractor and Classifier for SUIT | Published in Solar Physics | Solar Physics, Volume 300, article number 89, (2025) | 10.1007/s11207-025-02504-1 | null | astro-ph.SR astro-ph.IM cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The Solar Ultraviolet Imaging Telescope(SUIT) onboard Aditya-L1 is an imager that observes the solar photosphere and chromosphere through observations in the wavelength range of 200-400 nm. A comprehensive understanding of the plasma and thermodynamic properties of chromospheric and photospheric morphological structures requires a large sample statistical study, necessitating the development of automatic feature detection methods. To this end, we develop the feature detection algorithm SPACE-SUIT: Solar Phenomena Analysis and Classification using Enhanced vision techniques for SUIT, to detect and classify the solar chromospheric features to be observed from SUIT's Mg II k filter. Specifically, we target plage regions, sunspots, filaments, and off-limb structures. SPACE uses YOLO, a neural network-based model to identify regions of interest. We train and validate SPACE using mock-SUIT images developed from Interface Region Imaging Spectrometer(IRIS) full-disk mosaic images in Mg II k line, while we also perform detection on Level-1 SUIT data. SPACE achieves an approximate precision of 0.788, recall 0.863 and MAP of 0.874 on the validation mock SUIT FITS dataset. Given the manual labeling of our dataset, we perform "self-validation" by applying statistical measures and Tamura features on the ground truth and predicted bounding boxes. We find the distributions of entropy, contrast, dissimilarity, and energy to show differences in the features. These differences are qualitatively captured by the detected regions predicted by SPACE and validated with the observed SUIT images, even in the absence of labeled ground truth. This work not only develops a chromospheric feature extractor but also demonstrates the effectiveness of statistical metrics and Tamura features for distinguishing chromospheric features, offering independent validation for future detection schemes. | [
{
"version": "v1",
"created": "Wed, 11 Dec 2024 18:08:06 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 18:01:19 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Seth",
"Pranava",
""
],
[
"Upendran",
"Vishal",
""
],
[
"Anand",
"Megha",
""
],
[
"Sarkar",
"Janmejoy",
""
],
[
"Roy",
"Soumya",
""
],
[
"Chaki",
"Priyadarshan",
""
],
[
"Chowdhury",
"Pratyay",
""
],
[
"Ghosh",
"Borishan",
""
],
[
"Tripathi",
"Durgesh",
""
]
]
|
2412.09488 | Benjamin Li\'egeois | Benjamin Li\'egeois, Ramasubramanian Chitra, Nicol\`o Defenu | Echoes and defects in the Calogero model | null | Phys. Rev. A 111, 053304 (2025) | 10.1103/PhysRevA.111.053304 | null | cond-mat.quant-gas cond-mat.stat-mech cond-mat.str-el nlin.SI quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we extend the study of the interplay between scaling symmetries and statistics to one-dimensional fluids by studying the Calogero model in a harmonic trap modulated through time. The latter harbors an interpretation in terms of free particles imbued with exclusion statistics and is an example of a scale invariant fluid in one-dimension displaying $\mathrm{SO(2,1)}$ dynamical symmetry preserved by harmonic traps. Taking advantage of the dynamical symmetry, two experimentally relevant drive protocols spanning both quasistatic and nonadiabatic regimes are investigated and universal signatures of the interactions and exclusion statistics are uncovered in the ground-state echo amplitude and closely related ground-state fidelity. In particular, under both periodic modulation and slow drive through the gapless point of the trap frequency, enhanced interactions and exclusion are shown to favor the proliferation of defects and to hinder their annihilation, which leads to a universal decrease of ground-state fidelities and echo amplitudes. We also show that increasing exclusion sparks a sharp suppression of the likelihood of intermediate echoes beyond those imposed by the commensurability of a periodic drive and the natural frequency of the trap. | [
{
"version": "v1",
"created": "Thu, 12 Dec 2024 17:38:20 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 14:15:57 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Liégeois",
"Benjamin",
""
],
[
"Chitra",
"Ramasubramanian",
""
],
[
"Defenu",
"Nicolò",
""
]
]
|
2412.10145 | Wladislaw Krinitsin | Wladislaw Krinitsin, Niklas Tausendpfund, Matteo Rizzi, Markus Heyl, Markus Schmitt | Roughening dynamics of interfaces in the two-dimensional quantum Ising model | 17 pages, 10 figures | Phys. Rev. Lett. 134, 240402 (2025) | 10.1103/9bsk-x9rw | null | quant-ph cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | The properties of interfaces are key to understand the physics of matter. However, the study of quantum interface dynamics has remained an outstanding challenge. Here, we use large-scale Tree Tensor Network simulations to identify the dynamical signature of an interface roughening transition within the ferromagnetic phase of the 2D quantum Ising model. For initial domain wall profiles we find extended prethermal plateaus for smooth interfaces, whereas above the roughening transition the domain wall decays quickly. Our results can be readily explored experimentally in Rydberg atomic systems. | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 13:57:23 GMT"
},
{
"version": "v2",
"created": "Wed, 21 May 2025 12:21:47 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Krinitsin",
"Wladislaw",
""
],
[
"Tausendpfund",
"Niklas",
""
],
[
"Rizzi",
"Matteo",
""
],
[
"Heyl",
"Markus",
""
],
[
"Schmitt",
"Markus",
""
]
]
|
2412.10346 | Igor Zaliznyak | Jooseop Lee, Greta L. Chappell, Ryan E. Baumbach, Ayman H. Said, and Igor A. Zaliznyak | Large Kohn anomaly and phonon collapse induced by charge density wave in UPt$_2$Si$_2$ | 6 pages, 4 figures plus 5-page supplementary with 16 supplementary figures | null | null | null | cond-mat.str-el | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using high-energy-resolution inelastic x-ray scattering, we observe anomalous softening and damping of the transverse acoustic phonon in UPt$_2$Si$_2$ as the system is cooled towards the charge density wave (CDW) transition temperature, T$_{\rm{CDW}}$. The phonon exhibits a marked Kohn-type anomaly around the CDW wave vector, Q$_{\rm{CDW}}$, and becomes overdamped within a finite momentum range already well above T$_{\rm{CDW}}$. The dispersion anomaly is consistent with potential Fermi surface nesting, which together with the extended phonon collapse indicate strong electron-phonon coupling. The transition temperature estimated from the phonon softening is markedly lower than T$_{\rm{CDW}}$, consistent with a primarily electronic instability rather than a phonon-driven transition. Our results establish UPt$_2$Si$_2$ as a prime example of a strongly correlated electron CDW system with exceptionally strong electron-phonon coupling driving phonon softening and collapse. | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 18:41:51 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 20:55:49 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Lee",
"Jooseop",
""
],
[
"Chappell",
"Greta L.",
""
],
[
"Baumbach",
"Ryan E.",
""
],
[
"Said",
"Ayman H.",
""
],
[
"Zaliznyak",
"Igor A.",
""
]
]
|
2412.10435 | Xin Dong | Xin Dong, Sen Jia, Ming Rui Wang, Yan Li, Zhenheng Yang, Bingfeng Deng, Hongyu Xiong | COEF-VQ: Cost-Efficient Video Quality Understanding through a Cascaded Multimodal LLM Framework | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Recently, with the emergence of recent Multimodal Large Language Model (MLLM) technology, it has become possible to exploit its video understanding capability on different classification tasks. In practice, we face the difficulty of huge requirements for GPU resource if we need to deploy MLLMs online. In this paper, we propose COEF-VQ, a novel cascaded MLLM framework designed to enhance video quality understanding on the short-video platform while optimizing computational efficiency. Our approach integrates an entropy-based pre-filtering stage, where a lightweight model assesses uncertainty and selectively filters cases before passing them to the more computationally intensive MLLM for final evaluation. By prioritizing high-uncertainty samples for deeper analysis, our framework significantly reduces GPU usage while maintaining the strong classification performance of a full MLLM deployment. To demonstrate the effectiveness of COEF-VQ, we deploy this new framework onto the video management platform (VMP) at the short-video platform, and perform a series of detailed experiments on two in-house tasks related to video quality understanding. We show that COEF-VQ leads to substantial performance gains from the offline evaluation in these two tasks and effectively enhances platform safety with limit resource consumption, significantly reducing inappropriate content video view rate by 9.9% in a online A/B test without affecting engagement. Post-launch monitoring confirmed sustained improvements, validating its real-world impact. | [
{
"version": "v1",
"created": "Wed, 11 Dec 2024 08:10:32 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 21:50:59 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Dong",
"Xin",
""
],
[
"Jia",
"Sen",
""
],
[
"Wang",
"Ming Rui",
""
],
[
"Li",
"Yan",
""
],
[
"Yang",
"Zhenheng",
""
],
[
"Deng",
"Bingfeng",
""
],
[
"Xiong",
"Hongyu",
""
]
]
|
2412.11074 | Ji Zhao | Baocai Yin, Ji Zhao, Huajie Jiang, Ningning Hou, Yongli Hu, Amin Beheshti, Ming-Hsuan Yang, Yuankai Qi | Adapter-Enhanced Semantic Prompting for Continual Learning | This work has been submitted to the IJCV for possible publication | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Continual learning (CL) enables models to adapt to evolving data streams. A major challenge of CL is catastrophic forgetting, where new knowledge will overwrite previously acquired knowledge. Traditional methods usually retain the past data for replay or add additional branches in the model to learn new knowledge, which has high memory requirements. In this paper, we propose a novel lightweight CL framework, Adapter-Enhanced Semantic Prompting (AESP), which integrates prompt tuning and adapter techniques. Specifically, we design semantic-guided prompts to enhance the generalization ability of visual features and utilize adapters to efficiently fuse the semantic information, aiming to learn more adaptive features for the continual learning task. Furthermore, to choose the right task prompt for feature adaptation, we have developed a novel matching mechanism for prompt selection. Extensive experiments on three CL datasets demonstrate that our approach achieves favorable performance across multiple metrics, showing its potential for advancing CL. | [
{
"version": "v1",
"created": "Sun, 15 Dec 2024 06:14:55 GMT"
},
{
"version": "v2",
"created": "Wed, 23 Apr 2025 14:44:27 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 03:00:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yin",
"Baocai",
""
],
[
"Zhao",
"Ji",
""
],
[
"Jiang",
"Huajie",
""
],
[
"Hou",
"Ningning",
""
],
[
"Hu",
"Yongli",
""
],
[
"Beheshti",
"Amin",
""
],
[
"Yang",
"Ming-Hsuan",
""
],
[
"Qi",
"Yuankai",
""
]
]
|
2412.11163 | Martin Bohnert | Martin Bohnert | The Fine interior of dilations of a rational polytope | 20 pages; typos corrected | null | null | null | math.CO math.AG | http://creativecommons.org/licenses/by-sa/4.0/ | A nondegenerate toric hypersurface of negative Kodaira dimension can be characterized by the empty Fine interior of its Newton polytope according to recent work by Victor Batyrev, where the Fine interior is the rational subpolytope consisting of all points which have an integral distance of at least 1 to all integral supporting hyperplanes of the Newton polytope. Moreover, we get more information in this situation if we can describe how the Fine interior behaves for dilations of the Newton polytope, e.g. if we can determine the smallest dilation with a non-empty Fine interior. Therefore, in this article we give a purely combinatorial description of the Fine interiors of all dilations of a rational polytope, which allows us in particular to compute this smallest dilation and to classify all lattice 3-polytopes with empty Fine interior, for which we have only one point as Fine interior of the smallest dilation with non-empty Fine interior. | [
{
"version": "v1",
"created": "Sun, 15 Dec 2024 12:13:59 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 05:50:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Bohnert",
"Martin",
""
]
]
|
2412.11554 | Sungdong Lee | Sungdong Lee, Joshua Bang, Youngrae Kim, Hyungwon Choi, Sang-Yun Oh, Joong-Ho Won | Learning Massive-scale Partial Correlation Networks in Clinical Multi-omics Studies with HP-ACCORD | 25 pages, 6 figures, preprint | null | null | null | stat.ML cs.LG math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | Graphical model estimation from multi-omics data requires a balance between statistical estimation performance and computational scalability. We introduce a novel pseudolikelihood-based graphical model framework that reparameterizes the target precision matrix while preserving the sparsity pattern and estimates it by minimizing an $\ell_1$-penalized empirical risk based on a new loss function. The proposed estimator maintains estimation and selection consistency in various metrics under high-dimensional assumptions. The associated optimization problem allows for a provably fast computation algorithm using a novel operator-splitting approach and communication-avoiding distributed matrix multiplication. A high-performance computing implementation of our framework was tested using simulated data with up to one million variables, demonstrating complex dependency structures similar to those found in biological networks. Leveraging this scalability, we estimated a partial correlation network from a dual-omic liver cancer data set. The co-expression network estimated from the ultrahigh-dimensional data demonstrated superior specificity in prioritizing key transcription factors and co-activators by excluding the impact of epigenetic regulation, thereby highlighting the value of computational scalability in multi-omic data analysis. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 08:38:02 GMT"
},
{
"version": "v2",
"created": "Fri, 20 Dec 2024 17:56:10 GMT"
},
{
"version": "v3",
"created": "Thu, 03 Jul 2025 09:46:23 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Lee",
"Sungdong",
""
],
[
"Bang",
"Joshua",
""
],
[
"Kim",
"Youngrae",
""
],
[
"Choi",
"Hyungwon",
""
],
[
"Oh",
"Sang-Yun",
""
],
[
"Won",
"Joong-Ho",
""
]
]
|
2412.11556 | Yuchen Fu | Yuchen Fu, Zifeng Cheng, Zhiwei Jiang, Zhonghui Wang, Yafeng Yin, Zhengliang Li, Qing Gu | Token Prepending: A Training-Free Approach for Eliciting Better Sentence Embeddings from LLMs | Accept to ACL 2025 (Oral). Code are available on https://github.com/fuyuchenIfyw/token_prepending.git | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extracting sentence embeddings from large language models (LLMs) is a promising direction, as LLMs have demonstrated stronger semantic understanding capabilities. Previous studies typically focus on prompt engineering to elicit sentence embeddings from LLMs by prompting the model to encode sentence information into the embedding of the last token. However, LLMs are mostly decoder-only models with causal attention and the earlier tokens in the sentence cannot attend to the latter tokens, resulting in biased encoding of sentence information and cascading effects on the final decoded token. To this end, we propose a novel Token Prepending (TP) technique that prepends each layer's decoded sentence embedding to the beginning of the sentence in the next layer's input, allowing earlier tokens to attend to the complete sentence information under the causal attention mechanism. The proposed TP technique is a plug-and-play and training-free technique, which means it can be seamlessly integrated with various prompt-based sentence embedding methods and autoregressive LLMs. Extensive experiments on various Semantic Textual Similarity (STS) tasks and downstream classification tasks demonstrate that our proposed TP technique can significantly improve the performance of existing prompt-based sentence embedding methods across different LLMs, while incurring negligible additional inference cost. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 08:42:00 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 06:24:49 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Fu",
"Yuchen",
""
],
[
"Cheng",
"Zifeng",
""
],
[
"Jiang",
"Zhiwei",
""
],
[
"Wang",
"Zhonghui",
""
],
[
"Yin",
"Yafeng",
""
],
[
"Li",
"Zhengliang",
""
],
[
"Gu",
"Qing",
""
]
]
|
2412.11746 | Pei Zheng | Pei Zheng, Yidian Chen, Danning Li, Mei Huang, Yuxin Liu | Non-equilibrium dynamics of Goldstone excitation from holography | 28 pages, 16 figures | JHEP07(2025)029 | 10.1007/JHEP07(2025)029 | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | By using the holographic approach, we investigate the interplay between the order parameter and Goldstone modes in the real-time dynamics of the chiral phase transition. By quenching the system to a different thermal bath and obtaining different kinds of initial states, we solve the real-time evolution of the system numerically. Our main focus is on studying far-from equilibrium dynamics of strongly-coupled system and universal scaling behaviors related to such dynamics. The most striking observation is that an additional prethermalization stage emerges at non-critical temperature after introducing the Goldstone modes, which is not reported in any previous studies. Some basic properties related to this additional prethermalization stage have been discussed in detail. More interestingly, we also report a new scaling relation describing non-equilibrium evolution at non-critical temperature. This additional universal behavior indicates the appearance of a non-thermal fixed point in the dynamical region. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 13:04:42 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Dec 2024 13:43:06 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Zheng",
"Pei",
""
],
[
"Chen",
"Yidian",
""
],
[
"Li",
"Danning",
""
],
[
"Huang",
"Mei",
""
],
[
"Liu",
"Yuxin",
""
]
]
|
2412.12054 | Jannis Bolik | Jannis Bolik, Thomas Hofmann | Optimality of Right-Invariant Priors | null | null | null | null | math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We discuss optimal prediction for families of probability distributions with a locally compact topological group structure. Right-invariant priors were previously shown to yield a posterior predictive distribution minimizing the worst-case Kullback-Leibler risk among all predictive procedures. However, the assumptions for the proof are so strong that they rarely hold in practice and it is unclear when the density functions used in the proof exist. Therefore, we provide a measure-theoretic proof using a more appropriate set of assumptions. As an application, we show a strong optimality result for next-sample prediction for multivariate normal distributions. We also numerically evaluate prediction with right-invariant priors against other objective priors. | [
{
"version": "v1",
"created": "Mon, 16 Dec 2024 18:25:45 GMT"
},
{
"version": "v2",
"created": "Tue, 17 Dec 2024 12:01:02 GMT"
},
{
"version": "v3",
"created": "Tue, 24 Dec 2024 16:12:34 GMT"
},
{
"version": "v4",
"created": "Sat, 04 Jan 2025 16:30:43 GMT"
},
{
"version": "v5",
"created": "Mon, 13 Jan 2025 14:13:30 GMT"
},
{
"version": "v6",
"created": "Mon, 28 Apr 2025 12:35:25 GMT"
},
{
"version": "v7",
"created": "Wed, 14 May 2025 15:51:29 GMT"
},
{
"version": "v8",
"created": "Thu, 03 Jul 2025 15:42:03 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Bolik",
"Jannis",
""
],
[
"Hofmann",
"Thomas",
""
]
]
|
2412.12179 | Steven Phillips PhD | Steven Phillips, Naotsugu Tsuchiya | Towards a (meta-)mathematical theory of consciousness: universal (mapping) properties of experience | 37 pages, 4 figures, modified abstract, expanded introduction, moved some introductory material new section (Section 2), added references | null | null | null | q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | Conscious experience permeates our daily lives, yet general consensus on a theory of consciousness remains elusive. In the face of such difficulty, an alternative strategy is to address a more general (meta-level) version of the problem for insights into the original problem at hand. Category theory was developed for this purpose, i.e. as an axiomatic (meta-)mathematical theory for comparison of mathematical structures, and so affords a (formally) formal approach towards a theory of consciousness. In this way, category theory is used for comparison with Information Integration Theory (IIT) as a supposed axiomatic theory of consciousness, which says that every conscious state involves six axiomatic properties: the IIT axioms for consciousness. All six axioms are shown to follow from the categorical notion of a universal mapping property: a unique-existence condition for all instances in the domain of interest. Accordingly, this categorical approach affords a formal basis for further development of a (meta-)mathematical theory of consciousness, whence the slogan, ``Consciousness is a universal property.'' | [
{
"version": "v1",
"created": "Fri, 13 Dec 2024 03:06:24 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 02:21:01 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Phillips",
"Steven",
""
],
[
"Tsuchiya",
"Naotsugu",
""
]
]
|
2412.12595 | T. M. Dunster | T. M. Dunster | Uniform asymptotic expansions for Bessel functions of imaginary order and their zeros | null | null | null | null | math.CA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bessel and modified Bessel functions of imaginary order $i\nu$ ($\nu >0$) are studied. Asymptotic expansions are derived as $\nu \to \infty$ that are uniformly valid in unbounded complex domains, with error bounds provided. Coupled with appropriate connection formulas the approximations are uniformly valid for all complex argument. The expansions are of two forms, Liouville-Green type expansions only involving elementary functions, and ones involving Airy functions that are valid at a turning point of the defining differential equation. The new results have coefficients and error bounds that are simpler than in prior expansions, and further are used to construct asymptotic expansions for the zeros of Bessel and modified Bessel functions of large imaginary order, these being uniformly valid without restriction on their size (small or large). | [
{
"version": "v1",
"created": "Tue, 17 Dec 2024 06:53:06 GMT"
},
{
"version": "v2",
"created": "Thu, 02 Jan 2025 03:55:55 GMT"
},
{
"version": "v3",
"created": "Fri, 10 Jan 2025 05:18:19 GMT"
},
{
"version": "v4",
"created": "Wed, 02 Jul 2025 22:22:40 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Dunster",
"T. M.",
""
]
]
|
2412.13259 | Ivan Medina | Ivan Medina, Ois\'in Culhane, Felix C. Binder, Gabriel T. Landi, and John Goold | Anomalous discharging of quantum batteries: the ergotropic Mpemba effect | 11 pages, 4 figures. Close to published version | Phys. Rev. Lett. 134, 220402 (2025) | 10.1103/PhysRevLett.134.220402 | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Anomalous thermal relaxation is ubiquitous in nonequilibrium statistical mechanics. An emblematic example of this is the Mpemba effect, where an initially ``hot'' system cools faster than an initially ``cooler'' one. This effect has recently been studied in a variety of different classical and quantum settings. In this Letter, we find a novel signature of the Mpemba effect in the context of quantum batteries. We identify situations where batteries in higher charge states can discharge faster than less charged states. Specifically, we consider a quantum battery encoded in a single bosonic mode that is charged using unitary Gaussian operations. We show that the ergotropy, used here as a dynamical indicator of the energy stored in the battery, can be recast as a phase space relative entropy between the system's state and the unitarily connected passive state, at each time. Our formalism allows us to compute the ergotropy analytically under dissipative dynamics and allows us to understand the conditions which give rise to a Mpemba effect. We also find situations where two batteries charged to the same value using different operations can discharge at different rates. | [
{
"version": "v1",
"created": "Tue, 17 Dec 2024 19:00:02 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 20:25:47 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Medina",
"Ivan",
""
],
[
"Culhane",
"Oisín",
""
],
[
"Binder",
"Felix C.",
""
],
[
"Landi",
"Gabriel T.",
""
],
[
"Goold",
"John",
""
]
]
|
2412.13295 | Florian Kaltenberger | Florian Kaltenberger, Tommaso Melodia, Irfan Ghauri, Michele Polese, Raymond Knopp, Tien Thinh Nguyen, Sakthivel Velumani, Davide Villa, Leonardo Bonati, Robert Schmidt, Sagar Arora, Mikel Irazabal, Navid Nikaein | Driving Innovation in 6G Wireless Technologies: The OpenAirInterface Approach | This work has been submitted to the Elsesvier Computer Networks Journal for possible publication | null | 10.1016/j.comnet.2025.111410 | Computer Networks, Volume 269, September 2025, 111410 | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The development of 6G wireless technologies is rapidly advancing, with the 3rd Generation Partnership Project (3GPP) entering the pre-standardization phase and aiming to deliver the first specifications by 2028. This paper explores the OpenAirInterface (OAI) project, an open-source initiative that plays a crucial role in the evolution of 5G and future 6G networks. OAI provides a comprehensive implementation of 3GPP and O-RAN compliant networks, including Radio Access Network (RAN), Core Network (CN), and software-defined User Equipment (UE) components. This paper details the history and evolution of OAI, its licensing model, and the various projects under its umbrella, such as RAN, the CN, and the Operations, Administration and Maintenance (OAM) projects. It also highlights the development methodology, Continuous Integration/Continuous Delivery (CI/CD) processes, and end-to-end systems powered by OAI. Furthermore, the paper discusses the potential of OAI for 6G research, focusing on spectrum, reflective intelligent surfaces, and Artificial Intelligence (AI)/Machine Learning (ML) integration. The open-source approach of OAI is emphasized as essential for tackling the challenges of 6G, fostering community collaboration, and driving innovation in next-generation wireless technologies. | [
{
"version": "v1",
"created": "Tue, 17 Dec 2024 19:48:09 GMT"
},
{
"version": "v2",
"created": "Mon, 06 Jan 2025 19:51:53 GMT"
},
{
"version": "v3",
"created": "Fri, 18 Apr 2025 20:37:02 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Kaltenberger",
"Florian",
""
],
[
"Melodia",
"Tommaso",
""
],
[
"Ghauri",
"Irfan",
""
],
[
"Polese",
"Michele",
""
],
[
"Knopp",
"Raymond",
""
],
[
"Nguyen",
"Tien Thinh",
""
],
[
"Velumani",
"Sakthivel",
""
],
[
"Villa",
"Davide",
""
],
[
"Bonati",
"Leonardo",
""
],
[
"Schmidt",
"Robert",
""
],
[
"Arora",
"Sagar",
""
],
[
"Irazabal",
"Mikel",
""
],
[
"Nikaein",
"Navid",
""
]
]
|
2412.13730 | Dong Xie | Dong Xie, Chunling Xu | Dispersive Qubit Readout of Temperature | 9pages,2figures | Phys. Rev. A 111, 062412 (2025) | 10.1103/PhysRevA.111.062412 | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Squeezed light can exponentially increase the signal-to-noise ratio (SNR) of dispersive qubit readout, especially using a combination of injected external squeezing (IES) and intracavity squeezing (ICS). We further investigate whether IES and ICS can also exponentially improve the measurement precision of temperature. In the case of fully thermalized qubits isolated from thermal bath, the measurement precision of temperature can be improved exponentially when the temperature or measurement time or the input photon number approaches 0. In general, thermal fluctuations prevent the action of squeezed light. When multiple qubits maintain interacting with the thermal bath, the Heisenberg scaling can be achieved if the loss rate of the cavity is large and the coupling between the qubit and the optical cavity is weak enough. In the meantime, IES can also further promote the improvement of the measurement precision of the temperature exponentially. | [
{
"version": "v1",
"created": "Wed, 18 Dec 2024 11:08:45 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Xie",
"Dong",
""
],
[
"Xu",
"Chunling",
""
]
]
|
2412.14171 | Jihan Yang | Jihan Yang, Shusheng Yang, Anjali W. Gupta, Rilyn Han, Li Fei-Fei, Saining Xie | Thinking in Space: How Multimodal Large Language Models See, Remember, and Recall Spaces | Project page: https://vision-x-nyu.github.io/thinking-in-space.github.io/ | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Humans possess the visual-spatial intelligence to remember spaces from sequential visual observations. However, can Multimodal Large Language Models (MLLMs) trained on million-scale video datasets also ``think in space'' from videos? We present a novel video-based visual-spatial intelligence benchmark (VSI-Bench) of over 5,000 question-answer pairs, and find that MLLMs exhibit competitive - though subhuman - visual-spatial intelligence. We probe models to express how they think in space both linguistically and visually and find that while spatial reasoning capabilities remain the primary bottleneck for MLLMs to reach higher benchmark performance, local world models and spatial awareness do emerge within these models. Notably, prevailing linguistic reasoning techniques (e.g., chain-of-thought, self-consistency, tree-of-thoughts) fail to improve performance, whereas explicitly generating cognitive maps during question-answering enhances MLLMs' spatial distance ability. | [
{
"version": "v1",
"created": "Wed, 18 Dec 2024 18:59:54 GMT"
},
{
"version": "v2",
"created": "Wed, 02 Jul 2025 21:00:36 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Yang",
"Jihan",
""
],
[
"Yang",
"Shusheng",
""
],
[
"Gupta",
"Anjali W.",
""
],
[
"Han",
"Rilyn",
""
],
[
"Fei-Fei",
"Li",
""
],
[
"Xie",
"Saining",
""
]
]
|
2412.14513 | Yingzhou Mou | Yingzhou Mou, Yukio Hayashi | Vulnerable Connectivity Caused by Local Communities in Spatial Networks | Modified from the first version | PLOS ONE, 20(7): e0327203 (2025) | 10.1371/journal.pone.0327203 | null | cs.SI physics.soc-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Local communities by concentration of nodes connected with short links are widely observed in spatial networks. However, how such structure affects robustness of connectivity against malicious attacks remains unclear. This study investigates the impact of local communities on the robustness by modeling planar infrastructure reveals that the robustness is weakened by strong local communities in spatial networks. These results highlight the potential of long-distance links in mitigating the negative effects of local community on the robustness. | [
{
"version": "v1",
"created": "Thu, 19 Dec 2024 04:26:11 GMT"
},
{
"version": "v2",
"created": "Thu, 30 Jan 2025 08:29:49 GMT"
},
{
"version": "v3",
"created": "Mon, 03 Feb 2025 06:45:17 GMT"
},
{
"version": "v4",
"created": "Wed, 23 Apr 2025 04:44:12 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Mou",
"Yingzhou",
""
],
[
"Hayashi",
"Yukio",
""
]
]
|
2412.15932 | Jiaozi Wang | Jiaozi Wang, Merlin F\"ullgraf, Jochen Gemmer | Estimate of equilibration times of quantum correlation functions in the thermodynamic limit based on Lanczos coefficients | 13 pages, 105 figures | Phys. Rev. Lett. 135, 010403 (2025) | 10.1103/ztrj-2tm6 | null | cond-mat.stat-mech quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the equilibration times $T_\text{eq}$ of local observables in quantum chaotic systems by considering their auto-correlation functions. Based on the recursion method, we suggest a scheme to estimate $T_\text{eq}$ from the corresponding Lanczos coefficients that is expected to hold in the thermodynamic limit. We numerically find that if the observable eventually shows smoothly growing Lanczos coefficients, a finite number of the former is sufficient for a reasonable estimate of the equilibration time. This implies that equilibration occurs on a realistic time scale much shorter than the life of the universe. The numerical findings are further supported by analytical arguments. | [
{
"version": "v1",
"created": "Fri, 20 Dec 2024 14:24:17 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 06:35:19 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Wang",
"Jiaozi",
""
],
[
"Füllgraf",
"Merlin",
""
],
[
"Gemmer",
"Jochen",
""
]
]
|
2412.16338 | Jussara Moreira | Gast\~ao A. Braga, Jussara M. Moreira, Ant\^onio Marcos da Silva, Camila F. Souza | A Renormalization Group Approach to Higher Order Corrections to the Decay of Solutions to Nonlinear Integral Equations | null | null | null | null | math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we employ the Renormalization Group (RG) method to study higher order corrections to the long-time asymptotics of a class of nonlinear integral equations with a generalized heat kernel and with time-dependent coefficients. This is a follow up of the papers \cite{bib:souz-brag-more-art,bib:souz-brag-more-marginal,bib:brag-more-silv}. | [
{
"version": "v1",
"created": "Fri, 20 Dec 2024 20:59:56 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 12:07:39 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Braga",
"Gastão A.",
""
],
[
"Moreira",
"Jussara M.",
""
],
[
"da Silva",
"Antônio Marcos",
""
],
[
"Souza",
"Camila F.",
""
]
]
|
2412.16391 | Einar Bjarki Gunnarsson | Einar Bjarki Gunnarsson and Benedikt Vilji Magn\'usson and Jasmine Foo | Optimal dosing of anti-cancer treatment under drug-induced plasticity | 52 pages, 11 figures | null | null | null | q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While cancer has traditionally been considered a genetic disease, mounting evidence indicates an important role for non-genetic (epigenetic) mechanisms. Common anti-cancer drugs have recently been observed to induce the adoption of non-genetic drug-tolerant cell states, thereby accelerating the evolution of drug resistance. This confounds conventional high-dose treatment strategies aimed at maximal tumor reduction, since high doses can simultaneously promote non-genetic resistance. In this work, we study optimal dosing of anti-cancer treatment under drug-induced cell plasticity. We show that the optimal dosing strategy steers the tumor to a fixed equilibrium composition between sensitive and tolerant cells, while precisely balancing the trade-off between cell kill and tolerance induction. The optimal equilibrium strategy ranges from applying a low dose continuously to applying the maximum dose intermittently, depending on the dynamics of tolerance induction. We finally discuss how our approach can be integrated with in vitro data to derive patient-specific treatment insights. | [
{
"version": "v1",
"created": "Fri, 20 Dec 2024 22:57:58 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 16:45:19 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Gunnarsson",
"Einar Bjarki",
""
],
[
"Magnússon",
"Benedikt Vilji",
""
],
[
"Foo",
"Jasmine",
""
]
]
|
2412.16537 | Yuntian Chen | Yuntian Chen, Zhanyong Tang, Tianpei Lu, Bingsheng Zhang, Zhiying Shi and Zheng Wang | Accelerating Private Large Transformers Inference through Fine-grained Collaborative Computation | 14 Pages (with 4 Pages appendix; 14 Figures) | null | 10.1109/TIFS.2025.3584639 | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Homomorphic encryption (HE) and secret sharing (SS) enable computations on encrypted data, providing significant privacy benefits for large transformer-based models (TBM) in sensitive sectors like medicine and finance. However, private TBM inference incurs significant costs due to the coarse-grained application of HE and SS. We present FASTLMPI, a new approach to accelerate private TBM inference through fine-grained computation optimization. Specifically, through the fine-grained co-design of homomorphic encryption and secret sharing, FASTLMPI achieves efficient protocols for matrix multiplication, SoftMax, LayerNorm, and GeLU. In addition, FASTLMPI introduces a precise segmented approximation technique for differentiable non-linear, improving its fitting accuracy while maintaining a low polynomial degree. Compared to solution BOLT (S&P'24), FASTLMPI shows a remarkable 54% to 64% decrease in runtime and an impressive 72.2% reduction in communication costs. | [
{
"version": "v1",
"created": "Sat, 21 Dec 2024 08:33:12 GMT"
},
{
"version": "v2",
"created": "Sat, 25 Jan 2025 14:10:34 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Chen",
"Yuntian",
""
],
[
"Tang",
"Zhanyong",
""
],
[
"Lu",
"Tianpei",
""
],
[
"Zhang",
"Bingsheng",
""
],
[
"Shi",
"Zhiying",
""
],
[
"Wang",
"Zheng",
""
]
]
|
2412.17500 | Junying Lu | Junying Lu, Yaojun Chen | Oriented Ramsey numbers of some sparse graphs | 12 pages | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $H$ be an oriented graph without directed cycle. The oriented Ramsey number of $H$, denoted by $\overrightarrow{r}(H)$, is the smallest integer $N$ such that every tournament on $N$ vertices contains a copy of $H$. Rosenfeld (JCT-B, 1974) conjectured that $\overrightarrow{r}(H)=|H|$ if $H$ is a cycle of sufficiently large order, which was confirmed for $|H|\geq 9$ by Zein recently, and so does if $H$ is a path. Note that $\overrightarrow{r}(H)=|H|$ implies any tournament contains $H$ as a spanning subdigraph, it is interesting to ask when $\overrightarrow{r}(H)=|H|$ for $H$ being a sparse oriented graph. S\'os (1986) conjectured this is true if $H$ is a directed path plus an additional edge containing the origin of the path as one end, which was confirmed by Petrovi\'{c} (JGT, 1988). In this paper, we show that $\overrightarrow{r}(H)=|H|$ for $H$ being an oriented graph obtained by identifying a vertex of an antidirected cycle with one end of a directed path. Some other oriented Ramsey numbers for oriented graphs with one cycle are also discussed. | [
{
"version": "v1",
"created": "Mon, 23 Dec 2024 11:57:31 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 13:13:50 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Lu",
"Junying",
""
],
[
"Chen",
"Yaojun",
""
]
]
|
2412.17545 | Martin Bohnert | Martin Bohnert | Lattice 3-polytopes of lattice width 2 and corresponding toric hypersurfaces | 16 pages, 3 figures; typos corrected, figures revised | null | null | null | math.AG math.CO | http://creativecommons.org/licenses/by-sa/4.0/ | The Kodaira dimension of a nondegenerate toric hypersurface can be computed from the dimension of the Fine interior of its Newton polytope according to recent work of Victor Batyrev, where the Fine interior of the Newton polytope is the subpolytope consisting of all points which have an integral distance of at least $1$ to all integral supporting hyperplanes. In particular, if we have a Fine interior of codimension $1$, then the hypersurface is of general type and the Newton polytope has lattice width $2$. In this article we study this situation for lattice $3$-polytopes and the corresponding surfaces of general type. In particular, we classify all $2$-dimensional Fine interiors of those lattice $3$-polytopes which have at most $40$ interior lattice points, thus obtaining many examples of surfaces of general type and genus at most $40$. | [
{
"version": "v1",
"created": "Mon, 23 Dec 2024 13:06:07 GMT"
},
{
"version": "v2",
"created": "Thu, 03 Jul 2025 05:48:15 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Bohnert",
"Martin",
""
]
]
|
2412.17926 | Marina Bastrakova | Andrey E. Schegolev, Marina V. Bastrakova, Michael A. Sergeev, Anastasia A. Maksimovskaya, Nikolay V. Klenov, Igor I. Soloviev | Contemporary implementations of spiking bio-inspired neural networks | 30 pages, 24 figures | Mesoscience & Nanotechnology, volume 1, issue 1, 01005 (2024) | 10.64214/jmsn.01.01005 | null | cs.NE cond-mat.mes-hall cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | The extensive development of the field of spiking neural networks has led to many areas of research that have a direct impact on people's lives. As the most bio-similar of all neural networks, spiking neural networks not only allow the solution of recognition and clustering problems (including dynamics), but also contribute to the growing knowledge of the human nervous system. Our analysis has shown that the hardware implementation is of great importance, since the specifics of the physical processes in the network cells affect their ability to simulate the neural activity of living neural tissue, the efficiency of certain stages of information processing, storage and transmission. This survey reviews existing hardware neuromorphic implementations of bio-inspired spiking networks in the "semiconductor", "superconductor" and "optical" domains. Special attention is given to the possibility of effective "hybrids" of different approaches | [
{
"version": "v1",
"created": "Mon, 23 Dec 2024 19:33:43 GMT"
}
]
| 2025-07-04T00:00:00 | [
[
"Schegolev",
"Andrey E.",
""
],
[
"Bastrakova",
"Marina V.",
""
],
[
"Sergeev",
"Michael A.",
""
],
[
"Maksimovskaya",
"Anastasia A.",
""
],
[
"Klenov",
"Nikolay V.",
""
],
[
"Soloviev",
"Igor I.",
""
]
]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.