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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2507.01706 | Dominik Itner | Dominik Itner, Dmitrij Dreiling, Hauke Gravenkamp, Bernd Henning, Carolin Birk | A modified Levenberg-Marquardt method for estimating the elastic material parameters of polymer waveguides using residuals between autocorrelated frequency responses | null | null | null | null | cs.CE | http://creativecommons.org/licenses/by/4.0/ | In this contribution, we address the estimation of the frequency-dependent elastic parameters of polymers in the ultrasound range, which is formulated as an inverse problem. This inverse problem is implemented as a nonlinear regression-type optimization problem, in which the simulation signals are fitted to the measurement signals. These signals consist of displacement responses in waveguides, focusing on hollow cylindrical geometries to enhance the simulation efficiency. To accelerate the optimization and reduce the number of model evaluations and wait times, we propose two novel methods. First, we introduce an adaptation of the Levenberg-Marquardt method derived from a geometrical interpretation of the least-squares optimization problem. Second, we introduce an improved objective function based on the autocorrelated envelopes of the measurement and simulation signals. Given that this study primarily relies on simulation data to quantify optimization convergence, we aggregate the expected ranges of realistic material parameters and derive their distributions to ensure the reproducibility of optimizations with proper measurements. We demonstrate the effectiveness of our objective function modification and step adaptation for various materials with isotropic material symmetry by comparing them with a state-of-the-art optimization method. In all cases, our method reduces the total number of model evaluations, thereby shortening the time to identify the material parameters. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:39:33 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Itner",
"Dominik",
""
],
[
"Dreiling",
"Dmitrij",
""
],
[
"Gravenkamp",
"Hauke",
""
],
[
"Henning",
"Bernd",
""
],
[
"Birk",
"Carolin",
""
]
] |
2507.01707 | Vilasini Tinnaneri Sreekanth | Vilasini Tinnaneri Sreekanth, Jean-Luc Starck, and Sandrine Codis | Generative modeling of convergence maps based on predicted one-point statistics | 9 pages, 9 figures | null | null | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | Context: Weak gravitational lensing is a key cosmological probe for current and future large-scale surveys. While power spectra are commonly used for analyses, they fail to capture non-Gaussian information from nonlinear structure formation, necessitating higher-order statistics and methods for efficient map generation. Aims: To develop an emulator that generates accurate convergence maps directly from an input power spectrum and wavelet l1-norm without relying on computationally intensive simulations. Methods: We use either numerical or theoretical predictions to construct convergence maps by iteratively adjusting wavelet coefficients to match target marginal distributions and their inter-scale dependencies, incorporating higher-order statistical information. Results: The resulting kappa maps accurately reproduce the input power spectrum and exhibit higher-order statistical properties consistent with the input predictions, providing an efficient tool for weak lensing analyses. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:39:57 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Sreekanth",
"Vilasini Tinnaneri",
""
],
[
"Starck",
"Jean-Luc",
""
],
[
"Codis",
"Sandrine",
""
]
] |
2507.01708 | Go Tokumitsu | Go Tokumitsu | Functional central limit theorems for the dynamic elephant random walk | null | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove functional central limit theorems for the dynamic elephant random walk in the $\sqrt{n}$ and $\sqrt{n\log n}$ orders, by applying the martingale convergence theorem and Karamata's theory of regular variation. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:39:59 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tokumitsu",
"Go",
""
]
] |
2507.01709 | Paul Freulon | Paul Freulon, Nikitas Georgakis and Victor Panaretos | Entropic optimal transport beyond product reference couplings: the Gaussian case on Euclidean space | 32 pages, 4 figures | null | null | null | math.ST stat.ML stat.TH | http://creativecommons.org/licenses/by/4.0/ | The optimal transport problem with squared Euclidean cost consists in finding a coupling between two input measures that maximizes correlation. Consequently, the optimal coupling is often singular with respect to Lebesgue measure. Regularizing the optimal transport problem with an entropy term yields an approximation called entropic optimal transport. Entropic penalties steer the induced coupling toward a reference measure with desired properties. For instance, when seeking a diffuse coupling, the most popular reference measures are the Lebesgue measure and the product of the two input measures. In this work, we study the case where the reference coupling is not necessarily assumed to be a product. We focus on the Gaussian case as a motivating paradigm, and provide a reduction of this more general optimal transport criterion to a matrix optimization problem. This reduction enables us to provide a complete description of the solution, both in terms of the primal variable and the dual variables. We argue that flexibility in terms of the reference measure can be important in statistical contexts, for instance when one has prior information, when there is uncertainty regarding the measures to be coupled, or to reduce bias when the entropic problem is used to estimate the un-regularized transport problem. In particular, we show in numerical examples that choosing a suitable reference plan allows to reduce the bias caused by the entropic penalty. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:40:21 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Freulon",
"Paul",
""
],
[
"Georgakis",
"Nikitas",
""
],
[
"Panaretos",
"Victor",
""
]
] |
2507.01710 | Paul Francis | Paul Francis, David Wagner | Towards Better Attribute Inference Vulnerability Measures | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The purpose of anonymizing structured data is to protect the privacy of individuals in the data while retaining the statistical properties of the data. An important class of attack on anonymized data is attribute inference, where an attacker infers the value of an unknown attribute of a target individual given knowledge of one or more known attributes. A major limitation of recent attribute inference measures is that they do not take recall into account, only precision. It is often the case that attacks target only a fraction of individuals, for instance data outliers. Incorporating recall, however, substantially complicates the measure, because one must determine how to combine recall and precision in a composite measure for both the attack and baseline. This paper presents the design and implementation of an attribute inference measure that incorporates both precision and recall. Our design also improves on how the baseline attribute inference is computed. In experiments using a generic best row match attack on moderately-anonymized microdata, we show that in over 25\% of the attacks, our approach correctly labeled the attack to be at risk while the prior approach incorrectly labeled the attack to be safe. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:41:08 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Francis",
"Paul",
""
],
[
"Wagner",
"David",
""
]
] |
2507.01711 | Mingfu Yan | Mingfu Yan and Jiancheng Huang and Yifan Liu and Shifeng Chen | Component Adaptive Clustering for Generalized Category Discovery | Accepted by IEEE ICME 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generalized Category Discovery (GCD) tackles the challenging problem of categorizing unlabeled images into both known and novel classes within a partially labeled dataset, without prior knowledge of the number of unknown categories. Traditional methods often rely on rigid assumptions, such as predefining the number of classes, which limits their ability to handle the inherent variability and complexity of real-world data. To address these shortcomings, we propose AdaGCD, a cluster-centric contrastive learning framework that incorporates Adaptive Slot Attention (AdaSlot) into the GCD framework. AdaSlot dynamically determines the optimal number of slots based on data complexity, removing the need for predefined slot counts. This adaptive mechanism facilitates the flexible clustering of unlabeled data into known and novel categories by dynamically allocating representational capacity. By integrating adaptive representation with dynamic slot allocation, our method captures both instance-specific and spatially clustered features, improving class discovery in open-world scenarios. Extensive experiments on public and fine-grained datasets validate the effectiveness of our framework, emphasizing the advantages of leveraging spatial local information for category discovery in unlabeled image datasets. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:41:30 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yan",
"Mingfu",
""
],
[
"Huang",
"Jiancheng",
""
],
[
"Liu",
"Yifan",
""
],
[
"Chen",
"Shifeng",
""
]
] |
2507.01712 | Xinle Tian | Xinle Tian, Matthew Nunes, Emiko Dupont, Shaunagh Downing, Freddie Lichtenstein, Matt Burns | Using Wavelet Domain Fingerprints to Improve Source Camera Identification | null | null | null | null | cs.CV eess.IV stat.AP | http://creativecommons.org/licenses/by/4.0/ | Camera fingerprint detection plays a crucial role in source identification and image forensics, with wavelet denoising approaches proving to be particularly effective in extracting sensor pattern noise (SPN). In this article, we propose a modification to wavelet-based SPN extraction. Rather than constructing the fingerprint as an image, we introduce the notion of a wavelet domain fingerprint. This avoids the final inversion step of the denoising algorithm and allows fingerprint comparisons to be made directly in the wavelet domain. As such, our modification streamlines the extraction and comparison process. Experimental results on real-world datasets demonstrate that our method not only achieves higher detection accuracy but can also significantly improve processing speed. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:43:24 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tian",
"Xinle",
""
],
[
"Nunes",
"Matthew",
""
],
[
"Dupont",
"Emiko",
""
],
[
"Downing",
"Shaunagh",
""
],
[
"Lichtenstein",
"Freddie",
""
],
[
"Burns",
"Matt",
""
]
] |
2507.01713 | Leonidas Anastasios Karageorgos | Nikos Irges and Leonidas Karageorgos | Energy momentum tensor correlators in $\phi^4$ theory II: The spin-two sector | 36 pages | null | null | null | hep-th hep-ph | http://creativecommons.org/licenses/by/4.0/ | We extend the computation of the C_T charge of the 2-point function of the Energy-Momentum Tensor to 4-loops. We show that C_T decomposes into two sectors, the conformal sector, which encodes the value of the central charge at fixed points and an RG-sector that contains logarithmic and constant corrections proportional to the beta-function. This latter constitutes the main new result of this work and is inaccessible via CFT methods alone. Furthermore, we demonstrate that C_T satisfies an eigenvalue-like equation analogous to that of the spin-0 charge, as discussed in part I, though with a different in general eigenvalue. Finally we present three possible applications. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:43:30 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Irges",
"Nikos",
""
],
[
"Karageorgos",
"Leonidas",
""
]
] |
2507.01714 | Kevin Innerebner | Kevin Innerebner and Franz M. Rohrhofer and Bernhard C. Geiger | B-PL-PINN: Stabilizing PINN Training with Bayesian Pseudo Labeling | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Training physics-informed neural networks (PINNs) for forward problems often suffers from severe convergence issues, hindering the propagation of information from regions where the desired solution is well-defined. Haitsiukevich and Ilin (2023) proposed an ensemble approach that extends the active training domain of each PINN based on i) ensemble consensus and ii) vicinity to (pseudo-)labeled points, thus ensuring that the information from the initial condition successfully propagates to the interior of the computational domain.
In this work, we suggest replacing the ensemble by a Bayesian PINN, and consensus by an evaluation of the PINN's posterior variance. Our experiments show that this mathematically principled approach outperforms the ensemble on a set of benchmark problems and is competitive with PINN ensembles trained with combinations of Adam and LBFGS. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:44:31 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Innerebner",
"Kevin",
""
],
[
"Rohrhofer",
"Franz M.",
""
],
[
"Geiger",
"Bernhard C.",
""
]
] |
2507.01715 | Aditya Tomar | Aditya Tomar, Rudra Murthy, Pushpak Bhattacharyya | Stereotype Detection as a Catalyst for Enhanced Bias Detection: A Multi-Task Learning Approach | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Bias and stereotypes in language models can cause harm, especially in sensitive areas like content moderation and decision-making. This paper addresses bias and stereotype detection by exploring how jointly learning these tasks enhances model performance. We introduce StereoBias, a unique dataset labeled for bias and stereotype detection across five categories: religion, gender, socio-economic status, race, profession, and others, enabling a deeper study of their relationship. Our experiments compare encoder-only models and fine-tuned decoder-only models using QLoRA. While encoder-only models perform well, decoder-only models also show competitive results. Crucially, joint training on bias and stereotype detection significantly improves bias detection compared to training them separately. Additional experiments with sentiment analysis confirm that the improvements stem from the connection between bias and stereotypes, not multi-task learning alone. These findings highlight the value of leveraging stereotype information to build fairer and more effective AI systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:46:00 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tomar",
"Aditya",
""
],
[
"Murthy",
"Rudra",
""
],
[
"Bhattacharyya",
"Pushpak",
""
]
] |
2507.01716 | Zhaochen Ding | Zhaochen Ding, Zheng Guo, Luyi Liu | Embedding a Praeger-Xu graph into a surface | 27 pages | null | null | null | math.CO math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Rotary maps (orientably regular maps) are highly symmetric graph embeddings on orientable surfaces. This paper classifies all rotary maps whose underlying graphs are Praeger-Xu graphs, denoted $\operatorname{C}(p,r,s)$, for any odd prime $p$ that does not divide $r$. Our main result establishes a one-to-one correspondence between the isomorphism classes of these maps and the multiplicity-free representations of the dihedral group $\operatorname{D}_{2r}$ over the finite field $\mathbb{F}_p$. This work extends a recent classification for the case where $p=2$. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:46:57 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Ding",
"Zhaochen",
""
],
[
"Guo",
"Zheng",
""
],
[
"Liu",
"Luyi",
""
]
] |
2507.01717 | Gopichand Kanumolu | Gopichand Kanumolu, Ashok Urlana, Charaka Vinayak Kumar, Bala Mallikarjunarao Garlapati | Agent Ideate: A Framework for Product Idea Generation from Patents Using Agentic AI | AgentScen Workshop, IJCAI 2025 | null | null | null | cs.AI cs.IR cs.LG cs.MA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Patents contain rich technical knowledge that can inspire innovative product ideas, yet accessing and interpreting this information remains a challenge. This work explores the use of Large Language Models (LLMs) and autonomous agents to mine and generate product concepts from a given patent. In this work, we design Agent Ideate, a framework for automatically generating product-based business ideas from patents. We experimented with open-source LLMs and agent-based architectures across three domains: Computer Science, Natural Language Processing, and Material Chemistry. Evaluation results show that the agentic approach consistently outperformed standalone LLMs in terms of idea quality, relevance, and novelty. These findings suggest that combining LLMs with agentic workflows can significantly enhance the innovation pipeline by unlocking the untapped potential of business idea generation from patent data. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:47:17 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Kanumolu",
"Gopichand",
""
],
[
"Urlana",
"Ashok",
""
],
[
"Kumar",
"Charaka Vinayak",
""
],
[
"Garlapati",
"Bala Mallikarjunarao",
""
]
] |
2507.01718 | Evandro V. L. de Mello | H\'ercules H. Santana and, E. V. L. de Mello | Common superconducting transition in under and overdoped cuprate superconductors | 6 pages, 4 figures submitted | null | null | null | cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | Underdoped cuprate superconductors are believed to be strongly correlated with electronic systems with small phase stiffness leading to a large phase fluctuation region is known as the pseudogap state. With increasing doping it is generally agree that they become Fermi liquid, rendering the end of the superconductivity due to the sufficiently large electronic screening. However, this scenario does not stand against a recent experiment\cite{OverJJ2022} that combined magnetic susceptibility and Scanning Tunnelling Microscopy (STM) which measured superconducting gaps and amplitudes amid charge inhomogeneity far beyond the critical doping $p_{\rm c} \approx 0.27$. We reproduced these results by calculating the localized superconducting amplitudes that emerge out of charge inhomogeneities, which forms a mesoscopic granular superconductor with an array of Josephson junctions, whose average couplings determine the critical temperature $T_{\rm c}$. The calculations agree with the experiments and both yield that underdoped and overdoped compounds have superconducting long-range order by the same mechanism. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:48:04 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"and",
"Hércules H. Santana",
""
],
[
"de Mello",
"E. V. L.",
""
]
] |
2507.01719 | Dorian Peters | Dorian Peters, Fernanda Espinoza, Marco da Re, Guido Ivetta, Luciana Benotti, Rafael A. Calvo | Towards culturally-appropriate conversational AI for health in the majority world: An exploratory study with citizens and professionals in Latin America | null | null | null | null | cs.HC cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There is justifiable interest in leveraging conversational AI (CAI) for health across the majority world, but to be effective, CAI must respond appropriately within culturally and linguistically diverse contexts. Therefore, we need ways to address the fact that current LLMs exclude many lived experiences globally. Various advances are underway which focus on top-down approaches and increasing training data. In this paper, we aim to complement these with a bottom-up locally-grounded approach based on qualitative data collected during participatory workshops in Latin America. Our goal is to construct a rich and human-centred understanding of: a) potential areas of cultural misalignment in digital health; b) regional perspectives on chatbots for health and c)strategies for creating culturally-appropriate CAI; with a focus on the understudied Latin American context. Our findings show that academic boundaries on notions of culture lose meaning at the ground level and technologies will need to engage with a broader framework; one that encapsulates the way economics, politics, geography and local logistics are entangled in cultural experience. To this end, we introduce a framework for 'Pluriversal Conversational AI for Health' which allows for the possibility that more relationality and tolerance, rather than just more data, may be called for. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:48:25 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Peters",
"Dorian",
""
],
[
"Espinoza",
"Fernanda",
""
],
[
"da Re",
"Marco",
""
],
[
"Ivetta",
"Guido",
""
],
[
"Benotti",
"Luciana",
""
],
[
"Calvo",
"Rafael A.",
""
]
] |
2507.01720 | Mark Saffman | J. Scott, H. M. Lim, U. Singla, Q. Meece, J. T. Choy, S. Kolkowitz, T. M. Graham, M. Saffman | Laser cooling and qubit measurements on a forbidden transition in neutral Cs atoms | 5 figures and supplemental material | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We experimentally demonstrate background-free, hyperfine-level-selective measurements of individual Cs atoms by simultaneous cooling to $5.3~\mu\rm K$ and imaging on the $6s_{1/2}\rightarrow 5d_{5/2}$ electric-quadrupole transition. We achieve hyperfine resolved detection with fidelity 0.9993(4) and atom retention of 0.9954(5), limited primarily by vacuum lifetime. Performing state measurements in a 3D cooling configuration enables repeated low loss measurements. A theoretical analysis of an extension of the demonstrated approach based on quenching of the excited state with an auxiliary field, identifies parameters for hyperfine-resolved measurements with a projected fidelity of $\sim 0.9995 $ in $\sim 60~\mu\rm s$. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:51:17 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Scott",
"J.",
""
],
[
"Lim",
"H. M.",
""
],
[
"Singla",
"U.",
""
],
[
"Meece",
"Q.",
""
],
[
"Choy",
"J. T.",
""
],
[
"Kolkowitz",
"S.",
""
],
[
"Graham",
"T. M.",
""
],
[
"Saffman",
"M.",
""
]
] |
2507.01721 | Zhongwen Zhang | Zhongwen Zhang, Yuri Boykov | Soft Self-labeling and Potts Relaxations for Weakly-Supervised Segmentation | published at CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider weakly supervised segmentation where only a fraction of pixels have ground truth labels (scribbles) and focus on a self-labeling approach optimizing relaxations of the standard unsupervised CRF/Potts loss on unlabeled pixels. While WSSS methods can directly optimize such losses via gradient descent, prior work suggests that higher-order optimization can improve network training by introducing hidden pseudo-labels and powerful CRF sub-problem solvers, e.g. graph cut. However, previously used hard pseudo-labels can not represent class uncertainty or errors, which motivates soft self-labeling. We derive a principled auxiliary loss and systematically evaluate standard and new CRF relaxations (convex and non-convex), neighborhood systems, and terms connecting network predictions with soft pseudo-labels. We also propose a general continuous sub-problem solver. Using only standard architectures, soft self-labeling consistently improves scribble-based training and outperforms significantly more complex specialized WSSS systems. It can outperform full pixel-precise supervision. Our general ideas apply to other weakly-supervised problems/systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:52:34 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zhang",
"Zhongwen",
""
],
[
"Boykov",
"Yuri",
""
]
] |
2507.01722 | Enrico Cassano | Enrico Cassano, Riccardo Renzulli, Andrea Bragagnolo, Marco Grangetto | When Does Pruning Benefit Vision Representations? | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Pruning is widely used to reduce the complexity of deep learning models, but its effects on interpretability and representation learning remain poorly understood. This paper investigates how pruning influences vision models across three key dimensions: (i) interpretability, (ii) unsupervised object discovery, and (iii) alignment with human perception. We first analyze different vision network architectures to examine how varying sparsity levels affect feature attribution interpretability methods. Additionally, we explore whether pruning promotes more succinct and structured representations, potentially improving unsupervised object discovery by discarding redundant information while preserving essential features. Finally, we assess whether pruning enhances the alignment between model representations and human perception, investigating whether sparser models focus on more discriminative features similarly to humans. Our findings also reveal the presence of sweet spots, where sparse models exhibit higher interpretability, downstream generalization and human alignment. However, these spots highly depend on the network architectures and their size in terms of trainable parameters. Our results suggest a complex interplay between these three dimensions, highlighting the importance of investigating when and how pruning benefits vision representations. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:57:49 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Cassano",
"Enrico",
""
],
[
"Renzulli",
"Riccardo",
""
],
[
"Bragagnolo",
"Andrea",
""
],
[
"Grangetto",
"Marco",
""
]
] |
2507.01723 | Xupeng Zhu | Xupeng Zhu, Fan Wang, Robin Walters, Jane Shi | SE(3)-Equivariant Diffusion Policy in Spherical Fourier Space | Accepted at ICML 2025 | null | null | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | Diffusion Policies are effective at learning closed-loop manipulation policies from human demonstrations but generalize poorly to novel arrangements of objects in 3D space, hurting real-world performance. To address this issue, we propose Spherical Diffusion Policy (SDP), an SE(3) equivariant diffusion policy that adapts trajectories according to 3D transformations of the scene. Such equivariance is achieved by embedding the states, actions, and the denoising process in spherical Fourier space. Additionally, we employ novel spherical FiLM layers to condition the action denoising process equivariantly on the scene embeddings. Lastly, we propose a spherical denoising temporal U-net that achieves spatiotemporal equivariance with computational efficiency. In the end, SDP is end-to-end SE(3) equivariant, allowing robust generalization across transformed 3D scenes. SDP demonstrates a large performance improvement over strong baselines in 20 simulation tasks and 5 physical robot tasks including single-arm and bi-manual embodiments. Code is available at https://github.com/amazon-science/Spherical_Diffusion_Policy. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:58:15 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zhu",
"Xupeng",
""
],
[
"Wang",
"Fan",
""
],
[
"Walters",
"Robin",
""
],
[
"Shi",
"Jane",
""
]
] |
2507.01724 | Theresa Eimer | Micha Henheik, Theresa Eimer, Marius Lindauer | Revisiting Learning Rate Control | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The learning rate is one of the most important hyperparameters in deep learning, and how to control it is an active area within both AutoML and deep learning research. Approaches for learning rate control span from classic optimization to online scheduling based on gradient statistics. This paper compares paradigms to assess the current state of learning rate control. We find that methods from multi-fidelity hyperparameter optimization, fixed-hyperparameter schedules, and hyperparameter-free learning often perform very well on selected deep learning tasks but are not reliable across settings. This highlights the need for algorithm selection methods in learning rate control, which have been neglected so far by both the AutoML and deep learning communities. We also observe a trend of hyperparameter optimization approaches becoming less effective as models and tasks grow in complexity, even when combined with multi-fidelity approaches for more expensive model trainings. A focus on more relevant test tasks and new promising directions like finetunable methods and meta-learning will enable the AutoML community to significantly strengthen its impact on this crucial factor in deep learning. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:58:38 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Henheik",
"Micha",
""
],
[
"Eimer",
"Theresa",
""
],
[
"Lindauer",
"Marius",
""
]
] |
2507.01725 | Steffen Wolf | Miriam J\"ager and Steffen Wolf | More sophisticated is not always better: comparison of similarity measures for unsupervised learning of pathways in biomolecular simulations | This preprint is the unedited version of a manuscript that has been sent to a peer-reviewed scientific journal for consideration as article. Copyright with the authors and the publisher after publication | null | null | null | cond-mat.soft physics.bio-ph physics.comp-ph q-bio.BM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Finding process pathways in molecular simulations such as the unbinding paths of small molecule ligands from their binding sites at protein targets in a set of trajectories via unsupervised learning approaches requires the definition of a suitable similarity measure between trajectories. We here evaluate the performance of four such measures with varying degree of sophistication, i.e., Euclidean and Wasserstein distances, Procrustes analysis and dynamical time warping, when analyzing trajectory data from two different biased simulation driving protocols in the form of constant velocity constraint targeted MD and steered MD. In a streptavidin-biotin benchmark system with known ground truth clusters, Wasserstein distances yielded the best clustering performance, closely followed by Euclidean distances, both being the most computationally efficient similarity measures. In a more complex A2a receptor-inhibitor system, however, the simplest measure, i.e., Euclidean distances, was sufficient to reveal meaningful and interpretable clusters. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 13:58:46 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Jäger",
"Miriam",
""
],
[
"Wolf",
"Steffen",
""
]
] |
2507.01726 | Simon Olsson | Hang Zou and Martin Rahm and Anton Frisk Kockum and Simon Olsson | Generative flow-based warm start of the variational quantum eigensolver | 20 pages; 8 figures | null | null | null | quant-ph physics.chem-ph stat.ML | http://creativecommons.org/licenses/by/4.0/ | Hybrid quantum-classical algorithms like the variational quantum eigensolver (VQE) show promise for quantum simulations on near-term quantum devices, but are often limited by complex objective functions and expensive optimization procedures. Here, we propose Flow-VQE, a generative framework leveraging conditional normalizing flows with parameterized quantum circuits to efficiently generate high-quality variational parameters. By embedding a generative model into the VQE optimization loop through preference-based training, Flow-VQE enables quantum gradient-free optimization and offers a systematic approach for parameter transfer, accelerating convergence across related problems through warm-started optimization. We compare Flow-VQE to a number of standard benchmarks through numerical simulations on molecular systems, including hydrogen chains, water, ammonia, and benzene. We find that Flow-VQE outperforms baseline optimization algorithms, achieving computational accuracy with fewer circuit evaluations (improvements range from modest to more than two orders of magnitude) and, when used to warm-start the optimization of new systems, accelerates subsequent fine-tuning by up to 50-fold compared with Hartree--Fock initialization. Therefore, we believe Flow-VQE can become a pragmatic and versatile paradigm for leveraging generative modeling to reduce the costs of variational quantum algorithms. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:00:37 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zou",
"Hang",
""
],
[
"Rahm",
"Martin",
""
],
[
"Kockum",
"Anton Frisk",
""
],
[
"Olsson",
"Simon",
""
]
] |
2507.01727 | Siyang Tang | Siyang Tang, Wen-Hua Chen, Cunjia Liu | Auto-optimization of Energy Generation for Wave Energy Converters with Active Learning | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper presents an auto-optimization control framework for wave energy converters (WECs) to maximize energy generation under unknown and changing ocean conditions. The proposed control framework consists of two levels. The high-level controller operating at a longer time scale aims to maximize the average energy generation over several wave periods. The generated Power Take-Off (PTO) profile as the reference for the low-level physical system to follow. The new auto-optimization process leverages the parameterization of the non-stationary operation condition in WECs, establishing the relationship between the average energy generation and the key design parameters of the PTO force subject to the unknown wave parameters. The high-level controller is designed based on the concept of Dual Control for Exploration and Exploitation (DCEE) to quickly learn the unknown wave parameters by actively probing the ocean condition, while generating the optimal PTO profile. During this process, the uncertainty of the estimated wave condition is quantified and embedded in the optimization cost function to enable active learning. Simulation results under unknown regular and irregular waves demonstrate the effectiveness and robustness of this novel auto-optimization WEC systems with active learning, outperforming model predictive control, extremum seeking and classic Bang-Bang control approaches. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:02:33 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tang",
"Siyang",
""
],
[
"Chen",
"Wen-Hua",
""
],
[
"Liu",
"Cunjia",
""
]
] |
2507.01728 | Hao Wei | Hao Wei and Wanli Ni and Wen Wang and Wenjun Xu and Dusit Niyato and Ping Zhang | Token Communication in the Era of Large Models: An Information Bottleneck-Based Approach | null | null | null | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This letter proposes UniToCom, a unified token communication paradigm that treats tokens as the fundamental units for both processing and wireless transmission. Specifically, to enable efficient token representations, we propose a generative information bottleneck (GenIB) principle, which facilitates the learning of tokens that preserve essential information while supporting reliable generation across multiple modalities. By doing this, GenIB-based tokenization is conducive to improving the communication efficiency and reducing computational complexity. Additionally, we develop $\sigma$-GenIB to address the challenges of variance collapse in autoregressive modeling, maintaining representational diversity and stability. Moreover, we employ a causal Transformer-based multimodal large language model (MLLM) at the receiver to unify the processing of both discrete and continuous tokens under the next-token prediction paradigm. Simulation results validate the effectiveness and superiority of the proposed UniToCom compared to baselines under dynamic channel conditions. By integrating token processing with MLLMs, UniToCom enables scalable and generalizable communication in favor of multimodal understanding and generation, providing a potential solution for next-generation intelligent communications. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:03:01 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Wei",
"Hao",
""
],
[
"Ni",
"Wanli",
""
],
[
"Wang",
"Wen",
""
],
[
"Xu",
"Wenjun",
""
],
[
"Niyato",
"Dusit",
""
],
[
"Zhang",
"Ping",
""
]
] |
2507.01729 | Sven Ullmann | Sven Ullmann, Tobias Ehring, Robin Herkert, Bernard Haasdonk | A trust-region framework for optimization using Hermite kernel surrogate models | 29 pages | null | null | null | math.NA cs.NA math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we present a trust-region optimization framework that employs Hermite kernel surrogate models. The method targets optimization problems with computationally demanding objective functions, for which direct optimization is often impractical due to expensive function evaluations. To address these challenges, we leverage a trust-region strategy, where the objective function is approximated by an efficient surrogate model within a local neighborhood of the current iterate. In particular, we construct the surrogate using Hermite kernel interpolation and define the trust-region based on bounds for the interpolation error. As mesh-free techniques, kernel-based methods are naturally suited for medium- to high-dimensional problems. Furthermore, the Hermite formulation incorporates gradient information, enabling precise gradient estimates that are crucial for many optimization algorithms. We prove that the proposed algorithm converges to a stationary point, and we demonstrate its effectiveness through numerical experiments, which illustrate the convergence behavior as well as the efficiency gains compared to direct optimization. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:03:11 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Ullmann",
"Sven",
""
],
[
"Ehring",
"Tobias",
""
],
[
"Herkert",
"Robin",
""
],
[
"Haasdonk",
"Bernard",
""
]
] |
2507.01731 | Ze Yun | Ze Yun | On nefness of the lowest piece of Hodge modules | Comments are welcome! | null | null | null | math.AG | http://creativecommons.org/licenses/by/4.0/ | We give degree lower bounds for quotient line bundles of the lowest piece of a Hodge module induced by a variation of Hodge structures outside a simple normal crossing divisor. We also give sufficient conditions on when the lowest piece is nef. Classical results say when X is a curve or monodromies are unipotent, the lowest piece is a nef vector bundle on X. In dimension greater than or equal to two, if the monodromies at infinity are not unipotent there are counter-examples where nefness fails. This note aims at giving an explanation of failure of nefness. Our result is a lower bound on degrees of quotient line bundles of the pullback of the lowest piece to curves, in terms of local monodromy and intersection numbers with the boundary divisors. The proof uses twisted Hodge modules, which appears naturally when specializing the lowest piece to divisors. We also give geometric examples to show that the lower bound can be achieved. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:04:38 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yun",
"Ze",
""
]
] |
2507.01732 | Yang Cao | Yang Cao, Hangyu Lin, Xinwei Sun, Yuan Yao | Gold after Randomized Sand: Model-X Split Knockoffs for Controlled Transformation Selection | null | null | null | null | stat.ME | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Controlling the False Discovery Rate (FDR) in variable selection is crucial for reproducibility and preventing over-selection, particularly with the increasing prevalence of predictive modeling. The Split Knockoff method, a recent extension of the canonical Knockoffs framework, offers finite-sample FDR control for selecting sparse transformations, finding applications across signal processing, economics, information technology, and the life sciences. However, its current formulation is limited to fixed design settings, restricting its use to linear models. The question of whether it can be generalized to random designs, thereby accommodating a broader range of models beyond the linear case -- similar to the Model-X Knockoff framework -- remains unanswered. A major challenge in addressing transformational sparsity within random design settings lies in reconciling the combination of a random design with a deterministic transformation. To overcome this limitation, we propose the Model-X Split Knockoff method. Our method achieves FDR control for transformation selection in random designs, bridging the gap between existing approaches. This is accomplished by introducing an auxiliary randomized design that interacts with both the existing random design and the deterministic transformation, enabling the construction of Model-X Split Knockoffs. Like the classical Model-X framework, our method provides provable finite-sample FDR control under known or accurately estimated covariate distributions, regardless of the conditional distribution of the response. Importantly, it guarantees at least the same selection power as Model-X Knockoffs when both are applicable. Empirical studies, including simulations and real-world applications to Alzheimer's disease imaging and university ranking analysis, demonstrate robust FDR control and suggest improved selection power over the original Model-X approach. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:04:39 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Cao",
"Yang",
""
],
[
"Lin",
"Hangyu",
""
],
[
"Sun",
"Xinwei",
""
],
[
"Yao",
"Yuan",
""
]
] |
2507.01733 | Eddye Bustamante | Eddye Bustamante, Jos\'e Jim\'enez Urrea, Jorge Mej\'ia | A note on the uniqueness properties of solutions for the Schr\"odinger-Korteweg de Vries system | null | null | null | null | math.AP | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this work we prove that if $(u_i,v_i)$, $i=1,2$, are smooth enough solutions of the coupled Schr\"odinger-Korteweg-de Vries system \begin{align*} \left. \begin{array}{rl} i u_t+\partial_x^2 u &\hspace{-2mm}=\beta uv - |u|^2 u,\\ \partial_t v + \partial_x^3 v &\hspace{-2mm}=\gamma \partial_x |u|^2-\frac12\partial_x (v^2) \end{array} \right\} \end{align*} with appropriate decay at infinity such that at two different times $t_0=0$ and $t_1=1$ satisfy that $$u_1(0)-u_2(0),u_1(1)-u_2(1),v_1(0)-v_2(0),v_1(1)-v_2(1)\in H^1(e^{ax^{2}}dx),$$ for $a>0$ big enough, then $u_1=u_2$ and $v_1=v_2$.
(Let us recall that $f\in H^1(e^{ax^{2}} dx)$ iff $f\in L^2(e^{ax^{2}}dx)$ and $\partial_x f\in L^2(e^{ax^{2}}dx)$). | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:05:52 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Bustamante",
"Eddye",
""
],
[
"Urrea",
"José Jiménez",
""
],
[
"Mejía",
"Jorge",
""
]
] |
2507.01734 | Oliver Wardas | Oliver Wardas, Florian Matthes | LLMs for Legal Subsumption in German Employment Contracts | PrePrint - ICAIL25, Chicago | null | null | null | cs.CL | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Legal work, characterized by its text-heavy and resource-intensive nature, presents unique challenges and opportunities for NLP research. While data-driven approaches have advanced the field, their lack of interpretability and trustworthiness limits their applicability in dynamic legal environments. To address these issues, we collaborated with legal experts to extend an existing dataset and explored the use of Large Language Models (LLMs) and in-context learning to evaluate the legality of clauses in German employment contracts. Our work evaluates the ability of different LLMs to classify clauses as "valid," "unfair," or "void" under three legal context variants: no legal context, full-text sources of laws and court rulings, and distilled versions of these (referred to as examination guidelines). Results show that full-text sources moderately improve performance, while examination guidelines significantly enhance recall for void clauses and weighted F1-Score, reaching 80\%. Despite these advancements, LLMs' performance when using full-text sources remains substantially below that of human lawyers. We contribute an extended dataset, including examination guidelines, referenced legal sources, and corresponding annotations, alongside our code and all log files. Our findings highlight the potential of LLMs to assist lawyers in contract legality review while also underscoring the limitations of the methods presented. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:07:54 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Wardas",
"Oliver",
""
],
[
"Matthes",
"Florian",
""
]
] |
2507.01735 | Kai Chen | Kai Chen, Ruiyuan Gao, Lanqing Hong, Hang Xu, Xu Jia, Holger Caesar, Dengxin Dai, Bingbing Liu, Dzmitry Tsishkou, Songcen Xu, Chunjing Xu, Qiang Xu, Huchuan Lu, Dit-Yan Yeung | ECCV 2024 W-CODA: 1st Workshop on Multimodal Perception and Comprehension of Corner Cases in Autonomous Driving | ECCV 2024. Workshop page: https://coda-dataset.github.io/w-coda2024/ | null | null | null | cs.CV cs.AI cs.CL cs.LG | http://creativecommons.org/licenses/by/4.0/ | In this paper, we present details of the 1st W-CODA workshop, held in conjunction with the ECCV 2024. W-CODA aims to explore next-generation solutions for autonomous driving corner cases, empowered by state-of-the-art multimodal perception and comprehension techniques. 5 Speakers from both academia and industry are invited to share their latest progress and opinions. We collect research papers and hold a dual-track challenge, including both corner case scene understanding and generation. As the pioneering effort, we will continuously bridge the gap between frontier autonomous driving techniques and fully intelligent, reliable self-driving agents robust towards corner cases. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:10:25 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Chen",
"Kai",
""
],
[
"Gao",
"Ruiyuan",
""
],
[
"Hong",
"Lanqing",
""
],
[
"Xu",
"Hang",
""
],
[
"Jia",
"Xu",
""
],
[
"Caesar",
"Holger",
""
],
[
"Dai",
"Dengxin",
""
],
[
"Liu",
"Bingbing",
""
],
[
"Tsishkou",
"Dzmitry",
""
],
[
"Xu",
"Songcen",
""
],
[
"Xu",
"Chunjing",
""
],
[
"Xu",
"Qiang",
""
],
[
"Lu",
"Huchuan",
""
],
[
"Yeung",
"Dit-Yan",
""
]
] |
2507.01736 | Siyang Wang | Yangxin Fu and Yan Jiang and Siyang Wang | An energy-based discontinuous Galerkin method for the wave equation with nonsmooth solutions | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We develop a stable and high-order accurate discontinuous Galerkin method for the second order wave equation, specifically designed to handle nonsmooth solutions. Our approach integrates the energy-based discontinuous Galerkin method with the oscillation-free technique to effectively suppress spurious oscillations near solution discontinuities. Both stability analysis and apriori error estimates are established for common choices of numerical fluxes. We present a series of numerical experiments to confirm the optimal convergence rates for smooth solutions and its robustness in maintaining oscillation-free behavior for nonsmooth solutions in wave equations without or with nonlinear source terms. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:13:40 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Fu",
"Yangxin",
""
],
[
"Jiang",
"Yan",
""
],
[
"Wang",
"Siyang",
""
]
] |
2507.01738 | Ming Dai | Ming Dai, Wenxuan Cheng, Jiang-jiang Liu, Sen Yang, Wenxiao Cai, Yanpeng Sun, Wankou Yang | DeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy | ICCV 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-sa/4.0/ | Referring Image Segmentation (RIS) is a challenging task that aims to segment objects in an image based on natural language expressions. While prior studies have predominantly concentrated on improving vision-language interactions and achieving fine-grained localization, a systematic analysis of the fundamental bottlenecks in existing RIS frameworks remains underexplored. To bridge this gap, we propose DeRIS, a novel framework that decomposes RIS into two key components: perception and cognition. This modular decomposition facilitates a systematic analysis of the primary bottlenecks impeding RIS performance. Our findings reveal that the predominant limitation lies not in perceptual deficiencies, but in the insufficient multi-modal cognitive capacity of current models. To mitigate this, we propose a Loopback Synergy mechanism, which enhances the synergy between the perception and cognition modules, thereby enabling precise segmentation while simultaneously improving robust image-text comprehension. Additionally, we analyze and introduce a simple non-referent sample conversion data augmentation to address the long-tail distribution issue related to target existence judgement in general scenarios. Notably, DeRIS demonstrates inherent adaptability to both non- and multi-referents scenarios without requiring specialized architectural modifications, enhancing its general applicability. The codes and models are available at https://github.com/Dmmm1997/DeRIS. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:14:35 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Dai",
"Ming",
""
],
[
"Cheng",
"Wenxuan",
""
],
[
"Liu",
"Jiang-jiang",
""
],
[
"Yang",
"Sen",
""
],
[
"Cai",
"Wenxiao",
""
],
[
"Sun",
"Yanpeng",
""
],
[
"Yang",
"Wankou",
""
]
] |
2507.01739 | Sue Shi | Sue Shi, Maximilian C. H\"ubl, Galien P. Grosjean, Carl P. Goodrich, Scott R. Waitukaitis | Electrostatics overcome acoustic collapse to assemble, adapt, and activate levitated matter | null | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Acoustic levitation provides a unique method for manipulating small particles as it completely evades effects from gravity, container walls, or physical handling. These advantages make it a tantalizing platform for studying complex phenomena in many-particle systems, save for one severe limitation -- particles suspended by sound interact via acoustic scattering forces, which cause them to merge into a single dense object. To overcome this "acoustic collapse", we have developed a strategy that combines acoustic levitation with controlled electrostatic charging to assemble, adapt, and activate complex, many-particle systems. The key idea is to introduce electrostatic repulsion, which renders a so-called "mermaid" potential where interactions are attractive at short range and repulsive at long range. By controlling the balance between attraction/repulsion, we are able to levitate fully expanded structures where all particles are separated, fully collapsed structures where they are all in contact, and hybrid ones consisting of both expanded and collapsed components. We find that fully collapsed and expanded structures are inherently stable, whereas hybrid ones exhibit transient stability governed by acoustically unstable dimers. Furthermore, we show how electrostatics allow us to adapt between configurations on the fly, either by quasistatic discharge or discrete up/down charge steps. Finally, we demonstrate how large structures experience selective energy pumping from the acoustic field -- thrusting some particles into motion while others remain stationary -- leading to complex dynamics including coupled rotations and oscillations. Our approach provides an easy-to-implement and easy-to-understand solution to the pervasive problem of acoustic collapse, while simultaneously providing new insights into the assembly and activation of many-particle systems with complex interactions. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:16:25 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Shi",
"Sue",
""
],
[
"Hübl",
"Maximilian C.",
""
],
[
"Grosjean",
"Galien P.",
""
],
[
"Goodrich",
"Carl P.",
""
],
[
"Waitukaitis",
"Scott R.",
""
]
] |
2507.01740 | Trung-Dung Hoang | Trung-Dung Hoang, Alceu Bissoto, Vihangkumar V. Naik, Tim Fl\"uhmann, Artemii Shlychkov, Jos\'e Garcia-Tirado, Lisa M. Koch | A Real-Time Digital Twin for Type 1 Diabetes using Simulation-Based Inference | null | null | null | null | cs.LG q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | Accurately estimating parameters of physiological models is essential to achieving reliable digital twins. For Type 1 Diabetes, this is particularly challenging due to the complexity of glucose-insulin interactions. Traditional methods based on Markov Chain Monte Carlo struggle with high-dimensional parameter spaces and fit parameters from scratch at inference time, making them slow and computationally expensive. In this study, we propose a Simulation-Based Inference approach based on Neural Posterior Estimation to efficiently capture the complex relationships between meal intake, insulin, and glucose level, providing faster, amortized inference. Our experiments demonstrate that SBI not only outperforms traditional methods in parameter estimation but also generalizes better to unseen conditions, offering real-time posterior inference with reliable uncertainty quantification. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:21:03 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Hoang",
"Trung-Dung",
""
],
[
"Bissoto",
"Alceu",
""
],
[
"Naik",
"Vihangkumar V.",
""
],
[
"Flühmann",
"Tim",
""
],
[
"Shlychkov",
"Artemii",
""
],
[
"Garcia-Tirado",
"José",
""
],
[
"Koch",
"Lisa M.",
""
]
] |
2507.01741 | Kshitij Khare | Hongqiang Sun and Kshitij Khare | High dimensional convergence rates for sparse precision estimators for matrix-variate data | null | null | null | null | math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In several applications, the underlying structure of the data allows for the samples to be organized into a matrix variate form. In such settings, the underlying row and column covariance matrices are fundamental quantities of interest. We focus our attention on two popular estimators that have been proposed in the literature: a penalized sparse estimator called SMGM and a heuristic sample covariance estimator. We establish convergence rates for these estimators in relevant high-dimensional settings, where the row and column dimensions of the matrix are allowed to increase with the sample size. We show that high-dimensional convergence rate analyses for the SMGM estimator in previous literature are incorrect. We discuss the critical errors in these proofs, and present a different and novel approach. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:21:46 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Sun",
"Hongqiang",
""
],
[
"Khare",
"Kshitij",
""
]
] |
2507.01742 | Antonio Arroyo Polonio | A. Arroyo-Polonio, C. Kehrig, J.M. V\'ilchez, J. Iglesias-P\'aramo, E. Perez-Montero, S. Duarte Puertas, J. Gallego, D. Reverte | New clues on the extended HeII ionization in IZw18 from GTC/MEGARA and JWST/MIRI | null | null | 10.3847/2041-8213/ade874 | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | IZw18 is one of the lowest-metallicity star-forming galaxies known at z$\sim$0, considered a unique local analogue of the first galaxies. The origin of its hard ionizing continuum, expected to be a common feature in the early Universe and traced by He\textsc{ii} emission lines, remains intensely debated and challenging to explain. Here we combine optical (GTC/MEGARA) and mid-infrared (JWST/MIRI) integral field spectroscopic observations for IZw18 to shed new light on the high-ionization phenomenon. This letter reports the first detection of the high-ionization [Ne\textsc{v}]14.32 $\mu$m line in IZw18. Its emission is spatially extended and coincident with the He\textsc{ii} peak, revealing the presence of highly energetic ionizing sources that surpass mechanisms previously proposed on the basis of He\textsc{ii} alone. Our kinematic studies highlight that the He\textsc{ii}$\lambda$4686-emitting gas displays higher velocity dispersions and a different velocity pattern compared to the H$\beta$ emission, suggesting the presence of energetic processes such as shocks or stellar-driven feedback. Additionally, integrated spectra show asymmetric blueshifted profiles in the He\textsc{ii}$\lambda$4686 line, possibly indicating \textbf{early-stage} stellar-driven outflows potentially facilitating future ionizing photon leakage. Our spatial analysis also reveals differences in structure between the emission of H$\beta$ and He\textsc{ii}$\lambda$4686, with the He\textsc{ii}$\lambda$4686 peak offset by a projected distance of 140 pc from the peak H$\beta$ emission. This indicates distinct locations for the most extreme ionizing sources compared to moderate ionizing sources. Our findings underscore the complex interplay of physical processes in extremely metal-poor environments with \textbf{high-ionized} gas, offering new insights into the conditions prevailing in the early galaxies. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:22:17 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Arroyo-Polonio",
"A.",
""
],
[
"Kehrig",
"C.",
""
],
[
"Vílchez",
"J. M.",
""
],
[
"Iglesias-Páramo",
"J.",
""
],
[
"Perez-Montero",
"E.",
""
],
[
"Puertas",
"S. Duarte",
""
],
[
"Gallego",
"J.",
""
],
[
"Reverte",
"D.",
""
]
] |
2507.01743 | Lorenzo Pucci | Lorenzo Pucci, Luca Arcangeloni and Andrea Giorgetti | Position and Velocity Estimation Accuracy in MIMO-OFDM ISAC Networks: A Fisher Information Analysis | 19 pages, 6 figures, 3 tables | null | null | null | eess.SP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Integrated sensing and communication (ISAC) is a core technology for future wireless networks, enabling high-resolution sensing and reliable data transmission within a unified radio platform. This paper develops a theoretical framework to assess the estimation accuracy of target position and velocity in heterogeneous orthogonal frequency division multiplexing (OFDM)-based ISAC networks with multiple cooperative and distributed multiple-input multiple-output (MIMO) base stations (BSs). Using Fisher information analysis, we first derive closed-form Cram\'er-Rao lower bounds (CRLBs) for target localization in single monostatic and bistatic configurations. We then analyze the benefits of BS cooperation by deriving CRLBs for joint position and velocity estimation in a general setting that encompasses multiple cooperating monostatic systems and multistatic networks with multiple transmitters (Txs) and receivers (Rxs). The influence of key system parameters, including the number of BSs, bandwidth, antenna array configuration, and network geometry, is systematically examined. Numerical results highlight the performance gains enabled by cooperative sensing and provide insights to guide the design of future ISAC systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:23:10 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Pucci",
"Lorenzo",
""
],
[
"Arcangeloni",
"Luca",
""
],
[
"Giorgetti",
"Andrea",
""
]
] |
2507.01744 | Benjamin Jin | Benjamin Jin and Grant Mair and Joanna M. Wardlaw and Maria del C. Vald\'es Hern\'andez | Calibrated Self-supervised Vision Transformers Improve Intracranial Arterial Calcification Segmentation from Clinical CT Head Scans | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Vision Transformers (ViTs) have gained significant popularity in the natural image domain but have been less successful in 3D medical image segmentation. Nevertheless, 3D ViTs are particularly interesting for large medical imaging volumes due to their efficient self-supervised training within the masked autoencoder (MAE) framework, which enables the use of imaging data without the need for expensive manual annotations. intracranial arterial calcification (IAC) is an imaging biomarker visible on routinely acquired CT scans linked to neurovascular diseases such as stroke and dementia, and automated IAC quantification could enable their large-scale risk assessment. We pre-train ViTs with MAE and fine-tune them for IAC segmentation for the first time. To develop our models, we use highly heterogeneous data from a large clinical trial, the third International Stroke Trial (IST-3). We evaluate key aspects of MAE pre-trained ViTs in IAC segmentation, and analyse the clinical implications. We show: 1) our calibrated self-supervised ViT beats a strong supervised nnU-Net baseline by 3.2 Dice points, 2) low patch sizes are crucial for ViTs for IAC segmentation and interpolation upsampling with regular convolutions is preferable to transposed convolutions for ViT-based models, and 3) our ViTs increase robustness to higher slice thicknesses and improve risk group classification in a clinical scenario by 46%. Our code is available online. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:23:22 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Jin",
"Benjamin",
""
],
[
"Mair",
"Grant",
""
],
[
"Wardlaw",
"Joanna M.",
""
],
[
"Hernández",
"Maria del C. Valdés",
""
]
] |
2507.01745 | Anna Szymusiak | Anna Szymusiak | Beyond morphophoricity: $s$-tight IC measurements in geometric generalised probabilistic theories | 11 pages, 2 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | The analysed in this paper new class of $s$-tight IC measurements contains both morphophoric measurements, preserving the geometry of the states space, and tight IC measurements, introduced nearly 20 years ago by Scott in the quantum case as optimal for the task of linear quantum tomography. By looking at the mathematical side of these classes we discover their common feature, which is also preserved in the broader class of $s$-tight IC measurements: a particularly elegant form of the formula that can be seen as the generalised form of the Urgleichung known from the QBist approach to quantum theory. In particular, the tight IC measurements are identified as the ones for which this generalised Urgleichung takes an exceptionally simple form. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:23:26 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Szymusiak",
"Anna",
""
]
] |
2507.01746 | Jean-Francois Donati | J.-F. Donati, P.I. Cristofari, C. Moutou, A. L'Heureux, N.J. Cook, E. Artigau, S.H.P. Alencar, E. Gaidos, A. Vidotto, P. Petit, A. Carmona, T. Ray, and the SPIRou science team | Six-yr SPIRou monitoring of the young planet-host dwarf AU Mic | A&A in press (12 pages + 8-page appendix, 15 figures, 6 tables) | null | null | null | astro-ph.SR astro-ph.EP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we revisit our spectropolarimetric and velocimetric analysis of the young M dwarf AU Mic based on data collected with SPIRou at the Canada-France-Hawaii telescope, over a monitoring period of 2041 d from 2019 to 2024. The longitudinal magnetic field, the small-scale magnetic field, and the differential temperature of AU Mic, derived from the unpolarized and circularly-polarized spectra, were clearly modulated with the stellar rotation period, with a pattern that evolved over time. The magnetic modeling with Zeeman-Doppler imaging provides a consistent description of the global field of AU Mic that agrees not only with the Least-Squares Deconvolved profiles of the circularly-polarized and unpolarized spectral lines, but also with the small-scale field measurements derived from the broadening of spectral lines, for each of the 11 subsets of the full data. We find that the large-scale field was mostly poloidal, with a dominant dipole component slightly tilted to the rotation axis which decreased from 1.4 to 1.1 kG before increasing at the end of the campaign. The average small-scale field followed a similar trend, decreasing from 2.8 to 2.6 kG then rising. The long-term magnetic evolution we report for AU Mic suggests that, if cyclic, the cycle period is significantly longer than 6 yr. From velocimetric data, we derived improved mass estimates for the two transiting planets, respectively equal to M_b = 6.3+2.5-1.8 M_earth and M_c = 11.6+3.3-2.7 M_earth, yielding very contrasting densities of 0.32+0.13-0.10 and 2.9+1.1-0.8 g/cm3, and a new 90% confidence upper limit of 4.9 M_earth for candidate planet d (period 12.7 d) suspected to induce the transit-timing variations of b and c. We also confirm our claim regarding candidate planet e orbiting with a period of 33.11+-0.06 d, albeit with a smaller mass of M_e = 21.1+5.4-4.3 M_earth. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:23:42 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Donati",
"J. -F.",
""
],
[
"Cristofari",
"P. I.",
""
],
[
"Moutou",
"C.",
""
],
[
"L'Heureux",
"A.",
""
],
[
"Cook",
"N. J.",
""
],
[
"Artigau",
"E.",
""
],
[
"Alencar",
"S. H. P.",
""
],
[
"Gaidos",
"E.",
""
],
[
"Vidotto",
"A.",
""
],
[
"Petit",
"P.",
""
],
[
"Carmona",
"A.",
""
],
[
"Ray",
"T.",
""
],
[
"team",
"the SPIRou science",
""
]
] |
2507.01747 | Nora Gourmelon | Nora Gourmelon, Marcel Dreier, Martin Mayr, Thorsten Seehaus, Dakota Pyles, Matthias Braun, Andreas Maier, Vincent Christlein | SSL4SAR: Self-Supervised Learning for Glacier Calving Front Extraction from SAR Imagery | in IEEE Transactions on Geoscience and Remote Sensing. arXiv admin note: text overlap with arXiv:2501.05281 | null | 10.1109/TGRS.2025.3580945 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Glaciers are losing ice mass at unprecedented rates, increasing the need for accurate, year-round monitoring to understand frontal ablation, particularly the factors driving the calving process. Deep learning models can extract calving front positions from Synthetic Aperture Radar imagery to track seasonal ice losses at the calving fronts of marine- and lake-terminating glaciers. The current state-of-the-art model relies on ImageNet-pretrained weights. However, they are suboptimal due to the domain shift between the natural images in ImageNet and the specialized characteristics of remote sensing imagery, in particular for Synthetic Aperture Radar imagery. To address this challenge, we propose two novel self-supervised multimodal pretraining techniques that leverage SSL4SAR, a new unlabeled dataset comprising 9,563 Sentinel-1 and 14 Sentinel-2 images of Arctic glaciers, with one optical image per glacier in the dataset. Additionally, we introduce a novel hybrid model architecture that combines a Swin Transformer encoder with a residual Convolutional Neural Network (CNN) decoder. When pretrained on SSL4SAR, this model achieves a mean distance error of 293 m on the "CAlving Fronts and where to Find thEm" (CaFFe) benchmark dataset, outperforming the prior best model by 67 m. Evaluating an ensemble of the proposed model on a multi-annotator study of the benchmark dataset reveals a mean distance error of 75 m, approaching the human performance of 38 m. This advancement enables precise monitoring of seasonal changes in glacier calving fronts. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:24:23 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Gourmelon",
"Nora",
""
],
[
"Dreier",
"Marcel",
""
],
[
"Mayr",
"Martin",
""
],
[
"Seehaus",
"Thorsten",
""
],
[
"Pyles",
"Dakota",
""
],
[
"Braun",
"Matthias",
""
],
[
"Maier",
"Andreas",
""
],
[
"Christlein",
"Vincent",
""
]
] |
2507.01748 | Lifeng Yin | Lifeng Yin, Xiaoqi Liu, Yongyong Li | Existence and concentration of nontrivial solutions for quasilinear Schr\"{o}dinger equation with indefinite potential | null | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper is concerned with the quasilinear Schr\"{o}dinger equation \begin{align*} -\Delta u+V(x)u+\frac{k}{2}\Delta(u^2)u=f(u)\quad \text{in}~~\mathbb{R}^N\text{,} \end{align*} where $N\geq 3$, $k>0$, $V\in C(\R)$ is an indefinite potential. Under structural conditions on the potential $V$ and the nonlinearity $f$, we establish the existence of a nontrivial solution through a combination of a local linking argument, Morse theory, and the Moser iteration. Moreover, if $f$ is odd, we obtain an unbounded sequence of nontrivial solutions via the symmetric Mountain Pass Theorem. Additionally, as $k\rightarrow0$, we analyze the concentration behavior of nontrivial solutions. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:26:40 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yin",
"Lifeng",
""
],
[
"Liu",
"Xiaoqi",
""
],
[
"Li",
"Yongyong",
""
]
] |
2507.01749 | Mucahit Cevik | Arash Dehghan and Mucahit Cevik and Merve Bodur and Bissan Ghaddar | Joint Matching and Pricing for Crowd-shipping with In-store Customers | null | null | null | null | cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper examines the use of in-store customers as delivery couriers in a centralized crowd-shipping system, targeting the growing need for efficient last-mile delivery in urban areas. We consider a brick-and-mortar retail setting where shoppers are offered compensation to deliver time-sensitive online orders. To manage this process, we propose a Markov Decision Process (MDP) model that captures key uncertainties, including the stochastic arrival of orders and crowd-shippers, and the probabilistic acceptance of delivery offers. Our solution approach integrates Neural Approximate Dynamic Programming (NeurADP) for adaptive order-to-shopper assignment with a Deep Double Q-Network (DDQN) for dynamic pricing. This joint optimization strategy enables multi-drop routing and accounts for offer acceptance uncertainty, aligning more closely with real-world operations. Experimental results demonstrate that the integrated NeurADP + DDQN policy achieves notable improvements in delivery cost efficiency, with up to 6.7\% savings over NeurADP with fixed pricing and approximately 18\% over myopic baselines. We also show that allowing flexible delivery delays and enabling multi-destination routing further reduces operational costs by 8\% and 17\%, respectively. These findings underscore the advantages of dynamic, forward-looking policies in crowd-shipping systems and offer practical guidance for urban logistics operators. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:27:32 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Dehghan",
"Arash",
""
],
[
"Cevik",
"Mucahit",
""
],
[
"Bodur",
"Merve",
""
],
[
"Ghaddar",
"Bissan",
""
]
] |
2507.01750 | Jose Lopez | Jose A. Lopez, Georg Stemmer, H\'ector Cordourier Maruri | Generalizable Detection of Audio Deepfakes | 8 pages, 3 figures | null | null | null | eess.AS cs.SD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present our comprehensive study aimed at enhancing the generalization capabilities of audio deepfake detection models. We investigate the performance of various pre-trained backbones, including Wav2Vec2, WavLM, and Whisper, across a diverse set of datasets, including those from the ASVspoof challenges and additional sources. Our experiments focus on the effects of different data augmentation strategies and loss functions on model performance. The results of our research demonstrate substantial enhancements in the generalization capabilities of audio deepfake detection models, surpassing the performance of the top-ranked single system in the ASVspoof 5 Challenge. This study contributes valuable insights into the optimization of audio models for more robust deepfake detection and facilitates future research in this critical area. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:28:11 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Lopez",
"Jose A.",
""
],
[
"Stemmer",
"Georg",
""
],
[
"Maruri",
"Héctor Cordourier",
""
]
] |
2507.01751 | Melanie Lebental | Chenni Xu, Aswathy Sundaresan, Nazire-Beg\"um Kazkal, Clement Lafargue, Lior Zarfaty, Li-Gang Wang, Ofek Birnholtz, Dominique Decanini, Melanie Lebental, and Patrick Sebbah | Black hole optical analogue: photon sphere microlasers | 15 pages | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | The bell-like ringdown of the gravitational field in the last stage of the merging of massive black holes is now routinely detected on earth by the last generation of gravitational wave detectors. Its spectrum is interpreted as a sum of damped sinusoidal vibrations of the spacetime in the vicinity of the black hole. These so-called quasinormal modes are currently the subject of extensive studies, yet, their true nature remains elusive. Here, we emulate, in the laboratory, genuine four-dimension black hole metrics by two-dimensional optical curved surfaces that preserve the features of lightlike geodesics. %We establish the analogy with gravitational waves radiated by relaxing black holes, and We analytically compute the optical quasinormal modes and show that they are confined around the photon sphere, the unstable region around a black hole where spacetime curvature traps light in circular orbits. By 3D-printing non-Euclidean dye-doped microcavities, we demonstrate lasing at the photon sphere with a mode profile that closely matches the analytical prediction. These results paves the way for observing astrophysical phenomena in tabletop setups and is likely to inspire innovative designs in photonics. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:29:03 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Xu",
"Chenni",
""
],
[
"Sundaresan",
"Aswathy",
""
],
[
"Kazkal",
"Nazire-Begüm",
""
],
[
"Lafargue",
"Clement",
""
],
[
"Zarfaty",
"Lior",
""
],
[
"Wang",
"Li-Gang",
""
],
[
"Birnholtz",
"Ofek",
""
],
[
"Decanini",
"Dominique",
""
],
[
"Lebental",
"Melanie",
""
],
[
"Sebbah",
"Patrick",
""
]
] |
2507.01752 | Ismail Labiad | Ismail Labiad, Mathurin Videau, Matthieu Kowalski, Marc Schoenauer, Alessandro Leite, Julia Kempe, Olivier Teytaud | Tuning without Peeking: Provable Privacy and Generalization Bounds for LLM Post-Training | null | null | null | null | cs.LG cs.AI cs.CL cs.CR | http://creativecommons.org/licenses/by/4.0/ | Gradient-based optimization is the workhorse of deep learning, offering efficient and scalable training via backpropagation. However, its reliance on large volumes of labeled data raises privacy and security concerns such as susceptibility to data poisoning attacks and the risk of overfitting. In contrast, black box optimization methods, which treat the model as an opaque function, relying solely on function evaluations to guide optimization, offer a promising alternative in scenarios where data access is restricted, adversarial risks are high, or overfitting is a concern. However, black box methods also pose significant challenges, including poor scalability to high-dimensional parameter spaces, as prevalent in large language models (LLMs), and high computational costs due to reliance on numerous model evaluations. This paper introduces BBoxER, an evolutionary black-box method for LLM post-training that induces an information bottleneck via implicit compression of the training data. Leveraging the tractability of information flow, we provide strong theoretical bounds on generalization, differential privacy, susceptibility to data poisoning attacks, and robustness to extraction attacks. BBoxER operates on top of pre-trained LLMs, offering a lightweight and modular enhancement suitable for deployment in restricted or privacy-sensitive environments, in addition to non-vacuous generalization guarantees. In experiments with LLMs, we demonstrate empirically that Retrofitting methods are able to learn, showing how a few iterations of BBoxER improve performance and generalize well on a benchmark of reasoning datasets. This positions BBoxER as an attractive add-on on top of gradient-based optimization. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:29:30 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Labiad",
"Ismail",
""
],
[
"Videau",
"Mathurin",
""
],
[
"Kowalski",
"Matthieu",
""
],
[
"Schoenauer",
"Marc",
""
],
[
"Leite",
"Alessandro",
""
],
[
"Kempe",
"Julia",
""
],
[
"Teytaud",
"Olivier",
""
]
] |
2507.01753 | Yash Kulkarni | Yash Kulkarni, Susheela Sharma, Omid Rezayof, Siddhartha Kapuria, Jordan P. Amadio, Mohsen Khadem, Maryam Tilton, and Farshid Alambeigi | Augmented Bridge Spinal Fixation: A New Concept for Addressing Pedicle Screw Pullout via a Steerable Drilling Robot and Flexible Pedicle Screws | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | To address the screw loosening and pullout limitations of rigid pedicle screws in spinal fixation procedures, and to leverage our recently developed Concentric Tube Steerable Drilling Robot (CT-SDR) and Flexible Pedicle Screw (FPS), in this paper, we introduce the concept of Augmented Bridge Spinal Fixation (AB-SF). In this concept, two connecting J-shape tunnels are first drilled through pedicles of vertebra using the CT-SDR. Next, two FPSs are passed through this tunnel and bone cement is then injected through the cannulated region of the FPS to form an augmented bridge between two pedicles and reinforce strength of the fixated spine. To experimentally analyze and study the feasibility of AB-SF technique, we first used our robotic system (i.e., a CT-SDR integrated with a robotic arm) to create two different fixation scenarios in which two J-shape tunnels, forming a bridge, were drilled at different depth of a vertebral phantom. Next, we implanted two FPSs within the drilled tunnels and then successfully simulated the bone cement augmentation process. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:31:33 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Kulkarni",
"Yash",
""
],
[
"Sharma",
"Susheela",
""
],
[
"Rezayof",
"Omid",
""
],
[
"Kapuria",
"Siddhartha",
""
],
[
"Amadio",
"Jordan P.",
""
],
[
"Khadem",
"Mohsen",
""
],
[
"Tilton",
"Maryam",
""
],
[
"Alambeigi",
"Farshid",
""
]
] |
2507.01754 | Jean-Francois Donati | J.-F. Donati, P.I. Cristofari, B. Klein, B. Finociety, C. Moutou | Full Stokes magnetometry of the active M dwarfs AU Mic and EV Lac with SPIRou | A&A in press (12 pages + 3-page appendix, 7 figures, 4 tables) | null | null | null | astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We report in this paper circularly and linearly polarized observations of the young active M dwarfs AU Mic and EV Lac with the near-infrared SPIRou spectropolarimeter at the Canada-France-Hawaii Telescope, collected from August to October 2023 over a few rotation cycles of both stars. Applying Least-Squares Deconvolution (LSD) to our spectra, we detected Zeeman signatures in circular (Stokes V) and linear (Stokes QU) polarization, and Zeeman broadening in unpolarized (Stokes I) LSD profiles, all exhibiting clear rotational modulation. Using the stellar surface tomographic technique of Zeeman-Doppler imaging on our sets of observations, along with a simple parametric description of how the small-scale and large-scale fields relate to each other, we recovered the magnetic topologies of AU Mic and EV Lac successively from LSD Stokes V, Stokes IV and Stokes IVQU profiles, to investigate how the reconstructed maps evolve as we provide more information, and ultimately infer reliable magnetic maps of both stars. We find that AU Mic hosts a fairly simple and mostly poloidal large-scale field aligned with the rotation axis within about 10deg, whereas that of EV Lac is more complex, stronger and less axisymmetric. Both stars feature intense small-scale fields, of about 4 kG for AU Mic and 6 kG for EV Lac when averaged over the whole stellar surface. Stokes QU Zeeman signatures allow one to reconstruct stellar magnetic fields more reliably, and are especially useful for stars with more complex fields and low vsini's like EV Lac. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:31:52 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Donati",
"J. -F.",
""
],
[
"Cristofari",
"P. I.",
""
],
[
"Klein",
"B.",
""
],
[
"Finociety",
"B.",
""
],
[
"Moutou",
"C.",
""
]
] |
2507.01756 | Peng Zheng | Peng Zheng, Junke Wang, Yi Chang, Yizhou Yu, Rui Ma, and Zuxuan Wu | Rethinking Discrete Tokens: Treating Them as Conditions for Continuous Autoregressive Image Synthesis | accepted by iccv 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in large language models (LLMs) have spurred interests in encoding images as discrete tokens and leveraging autoregressive (AR) frameworks for visual generation. However, the quantization process in AR-based visual generation models inherently introduces information loss that degrades image fidelity. To mitigate this limitation, recent studies have explored to autoregressively predict continuous tokens. Unlike discrete tokens that reside in a structured and bounded space, continuous representations exist in an unbounded, high-dimensional space, making density estimation more challenging and increasing the risk of generating out-of-distribution artifacts. Based on the above findings, this work introduces DisCon (Discrete-Conditioned Continuous Autoregressive Model), a novel framework that reinterprets discrete tokens as conditional signals rather than generation targets. By modeling the conditional probability of continuous representations conditioned on discrete tokens, DisCon circumvents the optimization challenges of continuous token modeling while avoiding the information loss caused by quantization. DisCon achieves a gFID score of 1.38 on ImageNet 256$\times$256 generation, outperforming state-of-the-art autoregressive approaches by a clear margin. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:33:52 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zheng",
"Peng",
""
],
[
"Wang",
"Junke",
""
],
[
"Chang",
"Yi",
""
],
[
"Yu",
"Yizhou",
""
],
[
"Ma",
"Rui",
""
],
[
"Wu",
"Zuxuan",
""
]
] |
2507.01757 | Vignesh Kannan | Vignesh Kannan, Charles Dorn, Ute Drechsler, Dennis M. Kochmann | Microscale architected materials for elastic wave guiding: Fabrication and dynamic characterization across length and time scales | null | null | null | null | physics.app-ph | http://creativecommons.org/licenses/by/4.0/ | We present an experimental protocol for the fabrication and characterization of scalable microarchitected elastic waveguides. Using silicon microfabrication techniques, we develop free-standing 2D truss-based architected waveguides with a maximum diameter of 80 mm, unit cells size of 100 micrometer, and minimum beam width of 5 micrometer, thus achieving scale separation. To characterize elastic wave propagation, we introduce a custom-built scanning optical pump-probe experiment that enables contactless excitation of elastic wave modes and full spatio-temporal reconstruction of wave propagation across hundreds of unit cells with sub-unit cell resolution. Results on periodic architectures show excellent agreement with finite element simulations and equivalent experimental data at larger length scales. Motivated by scalable computational inverse design, we fabricate a specific example of a spatially graded waveguide and demonstrate its ability to guide elastic waves along an arbitrary pre-designed path. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:34:45 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Kannan",
"Vignesh",
""
],
[
"Dorn",
"Charles",
""
],
[
"Drechsler",
"Ute",
""
],
[
"Kochmann",
"Dennis M.",
""
]
] |
2507.01758 | Josey Stevens | Josey Stevens, Sebastian Deffner | Hamiltonian quantum gates -- energetic advantage from entangleability | 7 Pages, 1 figure | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Hamiltonian quantum gates controlled by classical electromagnetic fields form the basis of any realistic model of quantum computers. In this letter, we derive a lower bound on the field energy required to implement such gates and relate this energy to the expected gate error. We study the entangleability (ability to entangle qubits) of Hamiltonians and highlight how this feature of quantum gates can provide a means for more energetically efficient computation. Ultimately, we show that a universal quantum computer can be realized with vanishingly low energetic requirements but at the expense of arbitrarily large complexity. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:36:19 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Stevens",
"Josey",
""
],
[
"Deffner",
"Sebastian",
""
]
] |
2507.01759 | Nour Elhouda Tellache | Nour ElHouda Tellache and Lydia Aoudia and Mourad Boudhar | Scheduling on identical machines with conflicts to minimize the mean flow time | null | null | null | null | cs.DM math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper addresses the problem of scheduling jobs on identical machines with conflict constraints, where certain jobs cannot be scheduled simultaneously on different machines. We focus on the case where conflicts can be represented by a simple undirected graph, and the objective is to minimize the mean flow time. We show that the problem is NP-hard even on two machines and two distinct processing times. For unit-time jobs, the problem becomes NP-hard when the number of machines increases to three. We also identify polynomial-time solvable cases for specific classes of conflict graphs. For the general problem, we propose mathematical models, lower bounds, and a genetic algorithm. We evaluate their performance through computational experiments on a wide range of instances derived from well-known benchmark instances in the literature. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:36:35 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tellache",
"Nour ElHouda",
""
],
[
"Aoudia",
"Lydia",
""
],
[
"Boudhar",
"Mourad",
""
]
] |
2507.01760 | Elliot Kaplan | Allen Gehret, Elliot Kaplan, Nigel Pynn-Coates | Dimension theory for the asymptotic couple of the field of logarithmic transseries | 38 pages, 4 figures | null | null | MPIM-Bonn-2025 | math.LO | http://creativecommons.org/licenses/by/4.0/ | In this paper we completely characterize all dimension functions on all models of the theory $T_{\log}$ of the asymptotic couple of the field of logarithmic transseries (Dimension Theorem). This is done by characterizing the "small" $1$-variable definable sets (Small Sets Theorem). As a byproduct, we show that $T_{\log}$ is d-minimal and does not eliminate imaginaries. Separately, we provide an abstract criterion for d-minimality, which we use to observe some new examples of d-minimal expansions of valued fields. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:39:10 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Gehret",
"Allen",
""
],
[
"Kaplan",
"Elliot",
""
],
[
"Pynn-Coates",
"Nigel",
""
]
] |
2507.01761 | Nicolas Salvy | Nicolas Salvy, Hugues Talbot and Bertrand Thirion | Enhanced Generative Model Evaluation with Clipped Density and Coverage | null | null | null | null | cs.LG cs.AI stat.ML | http://creativecommons.org/licenses/by/4.0/ | Although generative models have made remarkable progress in recent years, their use in critical applications has been hindered by their incapacity to reliably evaluate sample quality. Quality refers to at least two complementary concepts: fidelity and coverage. Current quality metrics often lack reliable, interpretable values due to an absence of calibration or insufficient robustness to outliers. To address these shortcomings, we introduce two novel metrics, Clipped Density and Clipped Coverage. By clipping individual sample contributions and, for fidelity, the radii of nearest neighbor balls, our metrics prevent out-of-distribution samples from biasing the aggregated values. Through analytical and empirical calibration, these metrics exhibit linear score degradation as the proportion of poor samples increases. Thus, they can be straightforwardly interpreted as equivalent proportions of good samples. Extensive experiments on synthetic and real-world datasets demonstrate that Clipped Density and Clipped Coverage outperform existing methods in terms of robustness, sensitivity, and interpretability for evaluating generative models. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:40:00 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Salvy",
"Nicolas",
""
],
[
"Talbot",
"Hugues",
""
],
[
"Thirion",
"Bertrand",
""
]
] |
2507.01762 | Huayi Wei | Dong Wang, Chunyu Chen and Huayi Wei | Global Energy Minimization for Simplex Mesh Optimization: A Radius Ratio Approach to Sliver Elimination | null | null | null | null | math.NA cs.NA math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The quality of simplex mesh is crucial for the stability and accuracy of numerical simulations in finite element analysis and computational geometry. However, the presence of sliver elements in 3D simplex mesh can severely impact the results. This paper presents a novel method based on a radius ratio energy function to optimize the quality of simplex mesh elements. This method can effectively eliminate sliver elements, thereby enhancing mesh quality.The gradient of the proposed energy function can be decomposed into a matrix-vector product. With minor processing, the matrix becomes symmetric positive definite, and this symmetric positive definite matrix can serve as a preconditioner to significantly accelerate the optimization process. Experimental results demonstrate that this method has significant advantages in eliminating sliver elements and improving mesh quality. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:45:35 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Wang",
"Dong",
""
],
[
"Chen",
"Chunyu",
""
],
[
"Wei",
"Huayi",
""
]
] |
2507.01764 | Matteo Di Cristofaro | Matteo Di Cristofaro | Data interference: emojis, homoglyphs, and issues of data fidelity in corpora and their results | Author submitted manuscript | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Tokenisation - "the process of splitting text into atomic parts" (Brezina & Timperley, 2017: 1) - is a crucial step for corpus linguistics, as it provides the basis for any applicable quantitative method (e.g. collocations) while ensuring the reliability of qualitative approaches. This paper examines how discrepancies in tokenisation affect the representation of language data and the validity of analytical findings: investigating the challenges posed by emojis and homoglyphs, the study highlights the necessity of preprocessing these elements to maintain corpus fidelity to the source data. The research presents methods for ensuring that digital texts are accurately represented in corpora, thereby supporting reliable linguistic analysis and guaranteeing the repeatability of linguistic interpretations. The findings emphasise the necessity of a detailed understanding of both linguistic and technical aspects involved in digital textual data to enhance the accuracy of corpus analysis, and have significant implications for both quantitative and qualitative approaches in corpus-based research. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:46:26 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Di Cristofaro",
"Matteo",
""
]
] |
2507.01765 | Sarina Meyer | Sarina Meyer, Ekaterina Kolos, Ngoc Thang Vu | First Steps Towards Voice Anonymization for Code-Switching Speech | accepted at Interspeech 2025 | null | null | null | eess.AS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The goal of voice anonymization is to modify an audio such that the true identity of its speaker is hidden. Research on this task is typically limited to the same English read speech datasets, thus the efficacy of current methods for other types of speech data remains unknown. In this paper, we present the first investigation of voice anonymization for the multilingual phenomenon of code-switching speech. We prepare two corpora for this task and propose adaptations to a multilingual anonymization model to make it applicable for code-switching speech. By testing the anonymization performance of this and two language-independent methods on the datasets, we find that only the multilingual system performs well in terms of privacy and utility preservation. Furthermore, we observe challenges in performing utility evaluations on this data because of its spontaneous character and the limited code-switching support by the multilingual speech recognition model. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:46:44 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Meyer",
"Sarina",
""
],
[
"Kolos",
"Ekaterina",
""
],
[
"Vu",
"Ngoc Thang",
""
]
] |
2507.01766 | Da Guan | Tianwei Hou, Da Guan, Xin Sun, Anna Li, Wenqiang Yi, Yuanwei Liu, Arumugam Nallanathan | Reconfigurable Intelligent Surface aided Integrated-Navigation-and-Communication in Urban Canyons: A Satellite Selection Approach | null | null | null | null | cs.IT eess.SP math.IT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This study investigates the application of a simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS)-aided medium-Earth-orbit (MEO) satellite network for providing both global positioning services and communication services in the urban canyons, where the direct satellite-user links are obstructed. Superposition coding (SC) and successive interference cancellation (SIC) techniques are utilized for the integrated navigation and communication (INAC) networks, and the composed navigation and communication signals are reflected or transmitted to ground users or indoor users located in urban canyons. To meet diverse application needs, navigation-oriented (NO)-INAC and communication-oriented (CO)-INAC have been developed, each tailored according to distinct power allocation factors. We then proposed two algorithms, namely navigation-prioritized-algorithm (NPA) and communication-prioritized-algorithm (CPA), to improve the navigation or communication performance by selecting the satellite with the optimized position dilution of precision (PDoP) or with the best channel gain. The effectiveness of the proposed STAR-RIS-aided INAC network is quantified by analyzing the positioning error for navigation services and by evaluating communication performance through achievable ergodic rate metrics. Our satellite selection approach indicates that: the positioning services at the urban canyon users can be completed with the aid of STAR-RIS. 2) Additionally, it is observed that while a single STAR-RIS array can extend the navigational link, it fails to serve users in indoor scenarios, highlighting a limitation in the current system design. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:46:58 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Hou",
"Tianwei",
""
],
[
"Guan",
"Da",
""
],
[
"Sun",
"Xin",
""
],
[
"Li",
"Anna",
""
],
[
"Yi",
"Wenqiang",
""
],
[
"Liu",
"Yuanwei",
""
],
[
"Nallanathan",
"Arumugam",
""
]
] |
2507.01767 | Marco Rodrigues | Dylan Possama\"i, Marco Rodrigues, Alexandros Saplaouras | Mind the jumps: when 2BSDEs meet semi-martingales | 107 pages | null | null | null | math.PR math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct an aggregated version of the value processes associated with stochastic control problems, where the criterion to optimise is given by solutions to semi-martingale backward stochastic differential equations (BSDEs). The results can be applied to control problems where the triplet of semi-martingale characteristics is controlled in a possibly non-dominated case or where uncertainty about the characteristics is present in the optimisation. The construction also provides a time-consistent system of fully nonlinear conditional expectations on the Skorokhod space. We find the semi-martingale decomposition of the value function and characterise it as the solution to a semi-martingale second-order BSDE. The generality we seek allows for the treatment of controlled diffusions, pure-jump processes, and discrete-time processes in a unified setting. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:47:07 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Possamaï",
"Dylan",
""
],
[
"Rodrigues",
"Marco",
""
],
[
"Saplaouras",
"Alexandros",
""
]
] |
2507.01768 | Anis Yusof | Anis Yusof, Yuancheng Liu, Niklaus Kang, Choon Meng Seah, Zhenkai Liang, Ee-Chien Chang | Signals and Symptoms: ICS Attack Dataset From Railway Cyber Range | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by/4.0/ | The prevalence of cyberattacks on Industrial Control Systems (ICS) has highlighted the necessity for robust security measures and incident response to protect critical infrastructure. This is prominent when Operational Technology (OT) systems undergo digital transformation by integrating with Information Technology (IT) systems to enhance operational efficiency, adaptability, and safety. To support analysts in staying abreast of emerging attack patterns, there is a need for ICS datasets that reflect indicators representative of contemporary cyber threats. To address this, we conduct two ICS cyberattack simulations to showcase the impact of trending ICS cyberattacks on a railway cyber range that resembles the railway infrastructure. The attack scenario is designed to blend trending attack trends with attack patterns observed from historical ICS incidents. The resulting evidence is collected as datasets, serving as an essential resource for cyberattack analysis. This captures key indicators that are relevant to the current threat landscape, augmenting the effectiveness of security systems and analysts to protect against ICS cyber threats. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:47:31 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yusof",
"Anis",
""
],
[
"Liu",
"Yuancheng",
""
],
[
"Kang",
"Niklaus",
""
],
[
"Seah",
"Choon Meng",
""
],
[
"Liang",
"Zhenkai",
""
],
[
"Chang",
"Ee-Chien",
""
]
] |
2507.01769 | Yuta Takahashi | Yuta Takahashi and Shin-ichiro Sakai | Distance-based Relative Orbital Transition for Satellite Swarm Array Deployment Under J2 Perturbation | null | null | null | null | cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents an autonomous guidance and control strategy for a satellite swarm that enables scalable distributed space structures for innovative science and business opportunities. The averaged $J_2$ orbital parameters that describe the drift and periodic orbital motion were derived along with their target values to achieve a distributed space structure in a decentralized manner. This enabled the design of a distance-based orbital stabilizer to ensure autonomous deployment into a monolithic formation of a coplanar equidistant configuration on a user-defined orbital plane. Continuous formation control was assumed to be achieved through fuel-free actuation, such as satellite magnetic field interaction and differential aerodynamic forces, thereby maintaining long-term formation stability without thruster usage. A major challenge for such actuation systems is the potential loss of control capability due to increasing inter-satellite distances resulting from unstable orbital dynamics, particularly for autonomous satellite swarms. To mitigate this risk, our decentralized deployment controller minimized drift distance during unexpected communication outages. As a case study, we consider the deployment of palm-sized satellites into a coplanar equidistant formation in a $J_2$-perturbed orbit. Moreover, centralized grouping strategies are presented. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:50:21 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Takahashi",
"Yuta",
""
],
[
"Sakai",
"Shin-ichiro",
""
]
] |
2507.01770 | Guanglu Zhang | Guanglu Zhang, Qihang Shan, Jonathan Cagan | GPU-based complete search for nonlinear minimization subject to bounds | 36 pages, 3 figures | null | null | null | math.NA cs.AI cs.DC cs.MS cs.NA math.OC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper introduces a GPU-based complete search method to enclose the global minimum of a nonlinear function subject to simple bounds on the variables. Using interval analysis, coupled with the computational power and architecture of GPU, the method iteratively rules out the regions in the search domain where the global minimum cannot exist and leaves a finite set of regions where the global minimum must exist. For effectiveness, because of the rigor of interval analysis, the method is guaranteed to enclose the global minimum of the nonlinear function even in the presence of rounding errors. For efficiency, the method employs a novel GPU-based single program, single data parallel programming style to circumvent major GPU performance bottlenecks, and a variable cycling technique is also integrated into the method to reduce computational cost when minimizing large-scale nonlinear functions. The method is validated by minimizing 10 multimodal benchmark test functions with scalable dimensions, including the well-known Ackley function, Griewank function, Levy function, and Rastrigin function. These benchmark test functions represent grand challenges of global optimization, and enclosing the guaranteed global minimum of these benchmark test functions with more than 80 dimensions has not been reported in the literature. Our method completely searches the feasible domain and successfully encloses the guaranteed global minimum of these 10 benchmark test functions with up to 10,000 dimensions using only one GPU in a reasonable computation time, far exceeding the reported results in the literature due to the unique method design and implementation based on GPU architecture. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:54:52 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zhang",
"Guanglu",
""
],
[
"Shan",
"Qihang",
""
],
[
"Cagan",
"Jonathan",
""
]
] |
2507.01771 | Keith LeGrand | G. Andrew Siciliano and Keith A. LeGrand and Jackson Kulik | Higher-Order Tensor-Based Deferral of Gaussian Splitting for Orbit Uncertainty Propagation | null | null | null | null | eess.SP math.PR | http://creativecommons.org/licenses/by/4.0/ | Accurate propagation of orbital uncertainty is essential for a range of applications within space domain awareness. Adaptive Gaussian mixture-based approaches offer tractable nonlinear uncertainty propagation through splitting mixands to increase resolution in areas of stronger nonlinearities, as well as by reducing mixands to prevent unnecessary computational effort. Recent work introduced principled heuristics that incorporate information from the system dynamics and initial uncertainty to determine optimal directions for splitting. This paper develops adaptive uncertainty propagation methods based on these robust splitting techniques. A deferred splitting algorithm tightly integrated with higher-order splitting techniques is proposed and shown to offer substantial gains in computational efficiency without sacrificing accuracy. Second-order propagation of mixand moments is also seen to improve accuracy while retaining significant computational savings from deferred splitting. Different immediate and deferred splitting methods are compared in three representative test cases, including a geostationary orbit, a Molniya orbit, and a periodic three-body orbit. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:56:15 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Siciliano",
"G. Andrew",
""
],
[
"LeGrand",
"Keith A.",
""
],
[
"Kulik",
"Jackson",
""
]
] |
2507.01772 | Gianmassimo Tasinato | Bill Atkins, Debika Chowdhury, Alisha Marriott-Best, Gianmassimo Tasinato | Inflationary magnetogenesis beyond slow-roll and its induced gravitational waves | 22 pages, 5 figures | null | null | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | The origin of magnetic fields observed on both astrophysical and cosmological scales is a compelling problem that has the potential to shed light on the early Universe. We analytically investigate inflationary magnetogenesis in scenarios where a brief departure from slow-roll inflation - akin to mechanisms proposed for primordial black hole formation - leads to enhanced magnetic field generation with a growing power spectrum. Focusing on the Ratra model, we derive an analytic bound on the growth of the magnetic field power spectrum in this context, showing that the spectral index can reach $d \ln {\cal P}_B / d \ln k = 4.75$ during the growth phase. This growth enables amplification from CMB-safe large-scale amplitudes to values of astrophysical relevance. We further compute the stochastic gravitational wave background sourced by the resulting magnetic fields, incorporating their rich spectral features. Under suitable conditions, the induced signal exhibits a characteristic frequency dependence and amplitude within reach of future gravitational wave observatories, providing a distinctive signature of this mechanism and a specific class of templates for upcoming gravitational wave searches. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:57:06 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Atkins",
"Bill",
""
],
[
"Chowdhury",
"Debika",
""
],
[
"Marriott-Best",
"Alisha",
""
],
[
"Tasinato",
"Gianmassimo",
""
]
] |
2507.01773 | Ruihuai Liang | Bo Yang, Ruihuai Liang, Weixin Li, Han Wang, Xuelin Cao, Zhiwen Yu, Samson Lasaulce, M\'erouane Debbah, Mohamed-Slim Alouini, H. Vincent Poor, Chau Yuen | Frontiers of Generative AI for Network Optimization: Theories, Limits, and Visions | null | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While interest in the application of generative AI (GenAI) in network optimization has surged in recent years, its rapid progress has often overshadowed critical limitations intrinsic to generative models that remain insufficiently examined in existing literature. This survey provides a comprehensive review and critical analysis of GenAI in network optimization. We focus on the two dominant paradigms of GenAI including generative diffusion models (GDMs) and large pre-trained models (LPTMs), and organize our discussion around a categorization we introduce, dividing network optimization problems into two primary formulations: one-shot optimization and Markov decision process (MDP). We first trace key works, including foundational contributions from the AI community, and categorize current efforts in network optimization. We also review frontier applications of GDMs and LPTMs in other networking tasks, providing additional context. Furthermore, we present theoretical generalization bounds for GDMs in both one-shot and MDP settings, offering insights into the fundamental factors affecting model performance. Most importantly, we reflect on the overestimated perception of GenAI's general capabilities and caution against the all-in-one illusion it may convey. We highlight critical limitations, including difficulties in constraint satisfying, limited concept understanding, and the inherent probabilistic nature of outputs. We also propose key future directions, such as bridging the gap between generation and optimization. Although they are increasingly integrated in implementations, they differ fundamentally in both objectives and underlying mechanisms, necessitating a deeper understanding of their theoretical connections. Ultimately, this survey aims to provide a structured overview and a deeper insight into the strengths, limitations, and potential of GenAI in network optimization. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 14:59:49 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yang",
"Bo",
""
],
[
"Liang",
"Ruihuai",
""
],
[
"Li",
"Weixin",
""
],
[
"Wang",
"Han",
""
],
[
"Cao",
"Xuelin",
""
],
[
"Yu",
"Zhiwen",
""
],
[
"Lasaulce",
"Samson",
""
],
[
"Debbah",
"Mérouane",
""
],
[
"Alouini",
"Mohamed-Slim",
""
],
[
"Poor",
"H. Vincent",
""
],
[
"Yuen",
"Chau",
""
]
] |
2507.01774 | Jeroen van den Brink | Susmita Changdar, Oleksandr Suvorov, Andrii Kuibarov, Setti Thirupathaiah, Grigoriy Shipunov, Saicharan Aswartham, Sabine Wurmehl, Iryna Kovalchuk, Klaus Koepernik, Carsten Timm, Bernd B\"uchner, Ion Cosma Fulga, Sergey Borisenko and Jeroen van den Brink | Topological nodal $i$-wave superconductivity in PtBi$_2$ | 3 figures embedded | null | null | null | cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | Most superconducting materials are well-understood and conventional in the sense that the pairs of electrons that cause the superconductivity by their condensation have the highest possible symmetry. Famous exceptions are the enigmatic high-$T_c$ cuprate superconductors. Nodes in their superconducting gap are the fingerprint of their unconventional character and imply superconducting pairing of $d$-wave symmetry. Here, using angle-resolved photoemission spectroscopy, we observe that the Weyl semimetal PtBi$_2$ harbors nodes in its superconducting gap, implying unconventional $i$-wave pairing symmetry. At temperatures below $10\,\mathrm{K}$, the superconductivity in PtBi$_2$ gaps out its topological surface states, the Fermi arcs, while its bulk states remain normal. The nodes in the superconducting gap that we observe are located exactly at the center of the Fermi arcs, and imply the presence of topologically protected Majorana cones around this locus in momentum space. From this, we infer theoretically that robust zero-energy Majorana flat bands emerge at surface step edges. This not only establishes PtBi$_2$ surfaces as unconventional, topological $i$-wave superconductors but also as a promising material platform in the ongoing effort to generate and manipulate Majorana bound states. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:00:47 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Changdar",
"Susmita",
""
],
[
"Suvorov",
"Oleksandr",
""
],
[
"Kuibarov",
"Andrii",
""
],
[
"Thirupathaiah",
"Setti",
""
],
[
"Shipunov",
"Grigoriy",
""
],
[
"Aswartham",
"Saicharan",
""
],
[
"Wurmehl",
"Sabine",
""
],
[
"Kovalchuk",
"Iryna",
""
],
[
"Koepernik",
"Klaus",
""
],
[
"Timm",
"Carsten",
""
],
[
"Büchner",
"Bernd",
""
],
[
"Fulga",
"Ion Cosma",
""
],
[
"Borisenko",
"Sergey",
""
],
[
"Brink",
"Jeroen van den",
""
]
] |
2507.01775 | Haitao Wang | Haitao Wang | A Deterministic Partition Tree and Applications | To appear in ESA 2025 | null | null | null | cs.CG cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we present a deterministic variant of Chan's randomized partition tree [Discret. Comput. Geom., 2012]. This result leads to numerous applications. In particular, for $d$-dimensional simplex range counting (for any constant $d \ge 2$), we construct a data structure using $O(n)$ space and $O(n^{1+\epsilon})$ preprocessing time, such that each query can be answered in $o(n^{1-1/d})$ time (specifically, $O(n^{1-1/d} / \log^{\Omega(1)} n)$ time), thereby breaking an $\Omega(n^{1-1/d})$ lower bound known for the semigroup setting. Notably, our approach does not rely on any bit-packing techniques. We also obtain deterministic improvements for several other classical problems, including simplex range stabbing counting and reporting, segment intersection detection, counting and reporting, ray-shooting among segments, and more. Similar to Chan's original randomized partition tree, we expect that additional applications will emerge in the future, especially in situations where deterministic results are preferred. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:00:59 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Wang",
"Haitao",
""
]
] |
2507.01776 | Yuxuan Yang | Yuxuan Yang | Human-Machine Collaboration-Guided Space Design: Combination of Machine Learning Models and Humanistic Design Concepts | null | null | null | null | cs.HC cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The integration of machine learning (ML) into spatial design holds immense potential for optimizing space utilization, enhancing functionality, and streamlining design processes. ML can automate tasks, predict performance outcomes, and tailor spaces to user preferences. However, the emotional, cultural, and aesthetic dimensions of design remain crucial for creating spaces that truly resonate with users-elements that ML alone cannot address. The key challenge lies in harmonizing data-driven efficiency with the nuanced, subjective aspects of design. This paper proposes a human-machine collaboration framework to bridge this gap. An effective framework should recognize that while ML enhances design efficiency through automation and prediction, it must be paired with human creativity to ensure spaces are emotionally engaging and culturally relevant. Human designers contribute intuition, empathy, and cultural insight, guiding ML-generated solutions to align with users' emotional and cultural needs. Additionally, we explore how various ML models can be integrated with human-centered design principles. These models can automate design generation and optimization, while human designers refine the outputs to ensure emotional resonance and aesthetic appeal. Through case studies in office and residential design, we illustrate how this framework fosters both creativity and cultural relevance. By merging ML with human creativity, spatial design can achieve a balance of efficiency and emotional impact, resulting in environments that are both functional and deeply human. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:01:51 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Yang",
"Yuxuan",
""
]
] |
2507.01777 | Despoina Dimakou | Rodrigo Alonso, Despoina Dimakou, Yunji Ha, Valentin V. Khoze | Charge quantisation, monopoles and emergent symmetry in the Standard Model and its embeddings | 74 pages, 20 figures, 16 tables | null | null | IPPP/25/42 | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | This work studies the connection of the global properties of the SM gauge group to 1-form discrete symmetries, the possible non-Abelian embeddings of the SM group, and electric and magnetic charge quantisation. Building on previous work, we introduce indexes to characterise the group choices, connect the concept of compositeness degree to emergent electric 1-form symmetry, introduce a new model to fill in the $p=1$ gap, and analyse the magnetic spectrum while connecting its UV and IR realisations. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:02:49 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Alonso",
"Rodrigo",
""
],
[
"Dimakou",
"Despoina",
""
],
[
"Ha",
"Yunji",
""
],
[
"Khoze",
"Valentin V.",
""
]
] |
2507.01778 | Sandeep Kumar | Vivek Tetarwal, Sandeep Kumar | A Hybrid Ensemble Learning Framework for Image-Based Solar Panel Classification | 6 pages | null | null | null | cs.IT cs.CV math.IT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The installation of solar energy systems is on the rise, and therefore, appropriate maintenance techniques are required to be used in order to maintain maximum performance levels. One of the major challenges is the automated discrimination between clean and dirty solar panels. This paper presents a novel Dual Ensemble Neural Network (DENN) to classify solar panels using image-based features. The suggested approach utilizes the advantages offered by various ensemble models by integrating them into a dual framework, aimed at improving both classification accuracy and robustness. The DENN model is evaluated in comparison to current ensemble methods, showcasing its superior performance across a range of assessment metrics. The proposed approach performs the best compared to other methods and reaches state-of-the-art accuracy on experimental results for the Deep Solar Eye dataset, effectively serving predictive maintenance purposes in solar energy systems. It reveals the potential of hybrid ensemble learning techniques to further advance the prospects of automated solar panel inspections as a scalable solution to real-world challenges. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:07:43 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tetarwal",
"Vivek",
""
],
[
"Kumar",
"Sandeep",
""
]
] |
2507.01779 | Yash Kulkarni | Daniyal Maroufi, Xinyuan Huang, Yash Kulkarni, Omid Rezayof, Susheela Sharma, Vaibhav Goggela, Jordan P. Amadio, Mohsen Khadem, and Farshid Alambeigi | S3D: A Spatial Steerable Surgical Drilling Framework for Robotic Spinal Fixation Procedures | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we introduce S3D: A Spatial Steerable Surgical Drilling Framework for Robotic Spinal Fixation Procedures. S3D is designed to enable realistic steerable drilling while accounting for the anatomical constraints associated with vertebral access in spinal fixation (SF) procedures. To achieve this, we first enhanced our previously designed concentric tube Steerable Drilling Robot (CT-SDR) to facilitate steerable drilling across all vertebral levels of the spinal column. Additionally, we propose a four-Phase calibration, registration, and navigation procedure to perform realistic SF procedures on a spine holder phantom by integrating the CT-SDR with a seven-degree-of-freedom robotic manipulator. The functionality of this framework is validated through planar and out-of-plane steerable drilling experiments in vertebral phantoms. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:07:46 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Maroufi",
"Daniyal",
""
],
[
"Huang",
"Xinyuan",
""
],
[
"Kulkarni",
"Yash",
""
],
[
"Rezayof",
"Omid",
""
],
[
"Sharma",
"Susheela",
""
],
[
"Goggela",
"Vaibhav",
""
],
[
"Amadio",
"Jordan P.",
""
],
[
"Khadem",
"Mohsen",
""
],
[
"Alambeigi",
"Farshid",
""
]
] |
2507.01780 | Kyle Miller | Eric Vin, Kyle A. Miller, and Daniel J. Fremont | LeanLTL: A unifying framework for linear temporal logics in Lean | 9 pages, 3 figures; for associated project files see https://github.com/UCSCFormalMethods/LeanLTL; to be published in LIPIcs for ITP '25 | null | null | null | cs.LO cs.PL | http://creativecommons.org/licenses/by/4.0/ | We propose LeanLTL, a unifying framework for linear temporal logics in Lean 4. LeanLTL supports reasoning about traces that represent either infinite or finite linear time. The library allows traditional LTL syntax to be combined with arbitrary Lean expressions, making it straightforward to define properties involving numerical or other types. We prove that standard flavors of LTL can be embedded in our framework. The library also provides automation for reasoning about LeanLTL formulas in a way that facilitates using Lean's existing tactics. Finally, we provide examples illustrating the utility of the library in reasoning about systems that come from applications. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:07:47 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Vin",
"Eric",
""
],
[
"Miller",
"Kyle A.",
""
],
[
"Fremont",
"Daniel J.",
""
]
] |
2507.01781 | Dalia Rodr\'iguez-Salas | Dalia Rodr\'iguez-Salas, Christian Riess | BranchNet: A Neuro-Symbolic Learning Framework for Structured Multi-Class Classification | 18 pages, 3 figures (with two images each) | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce BranchNet, a neuro-symbolic learning framework that transforms decision tree ensembles into sparse, partially connected neural networks. Each branch, defined as a decision path from root to a parent of leaves, is mapped to a hidden neuron, preserving symbolic structure while enabling gradient-based optimization. The resulting models are compact, interpretable, and require no manual architecture tuning. Evaluated on a suite of structured multi-class classification benchmarks, BranchNet consistently outperforms XGBoost in accuracy, with statistically significant gains. We detail the architecture, training procedure, and sparsity dynamics, and discuss the model's strengths in symbolic interpretability as well as its current limitations, particularly on binary tasks where further adaptive calibration may be beneficial. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:07:58 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Rodríguez-Salas",
"Dalia",
""
],
[
"Riess",
"Christian",
""
]
] |
2507.01782 | Yinghui Ye | Yinghui Ye, Shuang Lu, Liqin Shi, Xiaoli Chu, and Sumei Sun | Symbiotic Backscatter Communication: A Design Perspective on the Modulation Scheme of Backscatter Devices | Submitted to IEEE Trans | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Symbiotic Backscatter Communication (SBC) has emerged as a spectrum-efficient and low-power communication technology, where backscatter devices (BDs) modulate and reflect incident radio frequency (RF) signals from primary transmitters (PTs). While previous studies have assumed a circularly symmetric complex Gaussian (CSCG) distribution for the BD's signal, this assumption may not be practical because the high complexity of generating CSCG signals is not supported by the low-cost BD. In this paper, we address this gap by investigating SBC for two low-complexity modulation schemes, i.e., $M$-ary amplitude-shift keying (MASK) and $M$-ary phase-shift keying (MPSK), where BD's signals inherently deviate from CSCG distribution. Our goal is to derive the achievable rate of the PT and BD under the MASK/MPSK and to design MASK/MPSK modulation scheme for maximizing the PT's rate. Towards this end, we first derive the expressions of both the PT's rate and BD's rate. Theoretical results reveal that whether or not the BD improves the PT's rate depends on the phase of MASK/MPSK modulation, while the BD's rate is independent of this phase. We then formulate two optimization problems to maximize the PT's rate by adjusting the phase under the MASK and MPSK modulation schemes, respectively, and derive the optimal phases for each modulation scheme in closed forms. Simulation results demonstrate that the optimal phase of MASK/MPSK can ensure an improvement in the PT's rate, and reveal that a low-order ASK modulation is better than a low-order PSK for the BD in terms of improving PT's rate, especially when the direct link is not significantly weaker than the backscatter link in SBC. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:07:59 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Ye",
"Yinghui",
""
],
[
"Lu",
"Shuang",
""
],
[
"Shi",
"Liqin",
""
],
[
"Chu",
"Xiaoli",
""
],
[
"Sun",
"Sumei",
""
]
] |
2507.01783 | Yu Zhang | Yu Zhang, Xin Sun, Tianwei Hou, Anna Li, Sofie Pollin, Yuanwei Liu, Arumugam Nallanathan | ASTARS empowered Satellite Positioning Approach for Urban Canyons and Indoor Environments | null | null | null | null | cs.IT math.IT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | To mitigate the loss of satellite navigation signals in urban canyons and indoor environments, we propose an active simultaneous transmitting and reflecting reconfigurable intelligent surface (ASTARS) empowered satellite positioning approach. Deployed on building structures, ASTARS reflects navigation signals to outdoor receivers in urban canyons and transmits signals indoors to bypass obstructions, providing high-precision positioning services to receivers in non-line-of-sight (NLoS) areas. The path between ASTARS and the receiver is defined as the extended line-of-sight (ELoS) path and an improved carrier phase observation equation is derived to accommodate that. The receiver compensates for its clock bias through network time synchronization, corrects the actual signal path distance to the satellite-to-receiver distance through a distance correction algorithm, and determines its position by using the least squares (LS) method. Mathematical modeling of the errors introduced by the proposed method is conducted, followed by simulation analysis to assess their impact. Simulation results show that: 1) in areas where GNSS signals are blocked, with time synchronization accuracy within a 10 ns error range, the proposed method provides positioning services with errors not exceeding 4 m for both indoor and outdoor receivers, outperforming conventional NLoS methods with positioning errors of more than 7 m; 2) the additional errors introduced by the proposed method do not exceed 3 m for time synchronization errors within 10 ns, which includes the phase shift, beamwidth error, time synchronization errors, and satellite distribution errors, outperforming traditional NLoS methods, which typically produce positioning errors greater than 5 m. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:08:37 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zhang",
"Yu",
""
],
[
"Sun",
"Xin",
""
],
[
"Hou",
"Tianwei",
""
],
[
"Li",
"Anna",
""
],
[
"Pollin",
"Sofie",
""
],
[
"Liu",
"Yuanwei",
""
],
[
"Nallanathan",
"Arumugam",
""
]
] |
2507.01784 | Saeed Nasseh | Saeed Nasseh, Maiko Ono, and Yuji Yoshino | Tor algebra of local rings with decomposable maximal ideal | null | null | null | null | math.AC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $(R,{\frak{m}}_R)$ be a commutative noetherian local ring. Assuming that ${\frak{m}}_R=$$I\oplus J$ is a direct sum decomposition, where $I$ and $J$ are non-zero ideals of $R$, we describe the structure of the Tor algebra of $R$ in terms of the Tor algebras of the rings $R/I$ and $R/J$. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:09:10 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Nasseh",
"Saeed",
""
],
[
"Ono",
"Maiko",
""
],
[
"Yoshino",
"Yuji",
""
]
] |
2507.01785 | Zhixun Chen | Zhixun Chen, Ping Guo, Wenhan Han, Yifan Zhang, Binbin Liu, Haobin Lin, Fengze Liu, Yan Zhao, Bingni Zhang, Taifeng Wang, Yin Zheng, Meng Fang | MuRating: A High Quality Data Selecting Approach to Multilingual Large Language Model Pretraining | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Data quality is a critical driver of large language model performance, yet existing model-based selection methods focus almost exclusively on English. We introduce MuRating, a scalable framework that transfers high-quality English data-quality signals into a single rater for 17 target languages. MuRating aggregates multiple English "raters" via pairwise comparisons to learn unified document-quality scores,then projects these judgments through translation to train a multilingual evaluator on monolingual, cross-lingual, and parallel text pairs. Applied to web data, MuRating selects balanced subsets of English and multilingual content to pretrain a 1.2 B-parameter LLaMA model. Compared to strong baselines, including QuRater, AskLLM, DCLM and so on, our approach boosts average accuracy on both English benchmarks and multilingual evaluations, with especially large gains on knowledge-intensive tasks. We further analyze translation fidelity, selection biases, and underrepresentation of narrative material, outlining directions for future work. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:11:12 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Chen",
"Zhixun",
""
],
[
"Guo",
"Ping",
""
],
[
"Han",
"Wenhan",
""
],
[
"Zhang",
"Yifan",
""
],
[
"Liu",
"Binbin",
""
],
[
"Lin",
"Haobin",
""
],
[
"Liu",
"Fengze",
""
],
[
"Zhao",
"Yan",
""
],
[
"Zhang",
"Bingni",
""
],
[
"Wang",
"Taifeng",
""
],
[
"Zheng",
"Yin",
""
],
[
"Fang",
"Meng",
""
]
] |
2507.01786 | Jord Nguyen | Jord Nguyen, Khiem Hoang, Carlo Leonardo Attubato, Felix Hofst\"atter | Probing Evaluation Awareness of Language Models | Technical AI Governance Workshop, ICML (Poster) | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | Language models can distinguish between testing and deployment phases -- a capability known as evaluation awareness. This has significant safety and policy implications, potentially undermining the reliability of evaluations that are central to AI governance frameworks and voluntary industry commitments. In this paper, we study evaluation awareness in Llama-3.3-70B-Instruct. We show that linear probes can separate real-world evaluation and deployment prompts, suggesting that current models internally represent this distinction. We also find that current safety evaluations are correctly classified by the probes, suggesting that they already appear artificial or inauthentic to models. Our findings underscore the importance of ensuring trustworthy evaluations and understanding deceptive capabilities. More broadly, our work showcases how model internals may be leveraged to support blackbox methods in safety audits, especially for future models more competent at evaluation awareness and deception. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:12:43 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Nguyen",
"Jord",
""
],
[
"Hoang",
"Khiem",
""
],
[
"Attubato",
"Carlo Leonardo",
""
],
[
"Hofstätter",
"Felix",
""
]
] |
2507.01787 | Marie-Therese Sekwenz | Marie-Therese Sekwenz, Ben Wagner, Hans De Bruijn | From Reports to Reality: Testing Consistency in Instagram's Digital Services Act Compliance Data | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by-sa/4.0/ | The Digital Services Act (DSA) introduces harmonized rules for content moderation and platform governance in the European Union, mandating robust compliance mechanisms, particularly for very large online platforms and search engines. This study examined compliance with DSA requirements, focusing on Instagram as a case study. We develop and apply a multi-level consistency framework to evaluate DSA compliance. Our findings contribute to the broader discussion on empirically-based regulation, providing insight into how researchers, regulators, auditors and platforms can better utilize DSA mechanisms to improve reporting and enforcement quality and accountability. This work underscores that consistency can help detect potential compliance failures. It also demonstrates that platforms should be evaluated as part of an interconnected ecosystem rather than through isolated processes, which is crucial for effective compliance evaluation under the DSA. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:13:25 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Sekwenz",
"Marie-Therese",
""
],
[
"Wagner",
"Ben",
""
],
[
"De Bruijn",
"Hans",
""
]
] |
2507.01788 | Montasir Shams | Montasir Shams, Chashi Mahiul Islam, Shaeke Salman, Phat Tran, Xiuwen Liu | Are Vision Transformer Representations Semantically Meaningful? A Case Study in Medical Imaging | 9 pages | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Vision transformers (ViTs) have rapidly gained prominence in medical imaging tasks such as disease classification, segmentation, and detection due to their superior accuracy compared to conventional deep learning models. However, due to their size and complex interactions via the self-attention mechanism, they are not well understood. In particular, it is unclear whether the representations produced by such models are semantically meaningful. In this paper, using a projected gradient-based algorithm, we show that their representations are not semantically meaningful and they are inherently vulnerable to small changes. Images with imperceptible differences can have very different representations; on the other hand, images that should belong to different semantic classes can have nearly identical representations. Such vulnerability can lead to unreliable classification results; for example, unnoticeable changes cause the classification accuracy to be reduced by over 60\%. %. To the best of our knowledge, this is the first work to systematically demonstrate this fundamental lack of semantic meaningfulness in ViT representations for medical image classification, revealing a critical challenge for their deployment in safety-critical systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:14:06 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Shams",
"Montasir",
""
],
[
"Islam",
"Chashi Mahiul",
""
],
[
"Salman",
"Shaeke",
""
],
[
"Tran",
"Phat",
""
],
[
"Liu",
"Xiuwen",
""
]
] |
2507.01789 | Shihan Zhang | Yunqing Huang, Shihan Zhang | The inverse source problem of stochastic wave equation | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | To address the ill-posedness of the inverse source problem for the one-dimensional stochastic Helmholtz equations without attenuation, this study develops a novel computational framework designed to mitigate this inherent challenge at the numerical implementation level. For the stochastic wave equation driven by a finite-jump L\'evy process (assuming that its jump amplitude obeys a Gaussian distribution and the jump time interval obeys a Poisson distribution), this paper firstly establish the existence of a mild solution to its direct problem satisfying a particular stability estimate. Building upon these theoretical foundations, we further investigate the well-posedness of the inverse problem and develop a methodology to reconstruct the unknown source terms $f$ and $g$ using the data of the wave field at the final time point $u(x,T)$. This work not only provides rigorous theoretical analysis and effective numerical schemes for solving inverse source problems in these two specific classes of stochastic wave equations, but also offers new perspectives and methodological approaches for addressing a broader range of wave propagation inverse problems characterized by non-Gaussian stochastic properties. The proposed framework demonstrates significant relevance for characterizing physical phenomena influenced by jump-type stochastic perturbations, offering promising applications in diverse domains including but not limited to seismic wave propagation analysis and financial market volatility modeling. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:14:20 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Huang",
"Yunqing",
""
],
[
"Zhang",
"Shihan",
""
]
] |
2507.01790 | Tianze Hua | Tianze Hua, Tian Yun, Ellie Pavlick | How Do Vision-Language Models Process Conflicting Information Across Modalities? | All code and resources are available at: https://github.com/ethahtz/vlm_conflicting_info_processing | null | null | null | cs.CL cs.AI cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | AI models are increasingly required to be multimodal, integrating disparate input streams into a coherent state representation on which subsequent behaviors and actions can be based. This paper seeks to understand how such models behave when input streams present conflicting information. Focusing specifically on vision-language models, we provide inconsistent inputs (e.g., an image of a dog paired with the caption "A photo of a cat") and ask the model to report the information present in one of the specific modalities (e.g., "What does the caption say / What is in the image?"). We find that models often favor one modality over the other, e.g., reporting the image regardless of what the caption says, but that different models differ in which modality they favor. We find evidence that the behaviorally preferred modality is evident in the internal representational structure of the model, and that specific attention heads can restructure the representations to favor one modality over the other. Moreover, we find modality-agnostic "router heads" which appear to promote answers about the modality requested in the instruction, and which can be manipulated or transferred in order to improve performance across datasets and modalities. Together, the work provides essential steps towards identifying and controlling if and how models detect and resolve conflicting signals within complex multimodal environments. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:15:14 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Hua",
"Tianze",
""
],
[
"Yun",
"Tian",
""
],
[
"Pavlick",
"Ellie",
""
]
] |
2507.01791 | Zihong Guo | Zihong Guo, Chen Wan, Yayin Zheng, Hailing Kuang, Xiaohai Lu | Boosting Adversarial Transferability Against Defenses via Multi-Scale Transformation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The transferability of adversarial examples poses a significant security challenge for deep neural networks, which can be attacked without knowing anything about them. In this paper, we propose a new Segmented Gaussian Pyramid (SGP) attack method to enhance the transferability, particularly against defense models. Unlike existing methods that generally focus on single-scale images, our approach employs Gaussian filtering and three types of downsampling to construct a series of multi-scale examples. Then, the gradients of the loss function with respect to each scale are computed, and their average is used to determine the adversarial perturbations. The proposed SGP can be considered an input transformation with high extensibility that is easily integrated into most existing adversarial attacks. Extensive experiments demonstrate that in contrast to the state-of-the-art methods, SGP significantly enhances attack success rates against black-box defense models, with average attack success rates increasing by 2.3% to 32.6%, based only on transferability. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:16:30 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Guo",
"Zihong",
""
],
[
"Wan",
"Chen",
""
],
[
"Zheng",
"Yayin",
""
],
[
"Kuang",
"Hailing",
""
],
[
"Lu",
"Xiaohai",
""
]
] |
2507.01792 | Peng Zheng | Peng Zheng, Ye Wang, Rui Ma, and Zuxuan Wu | FreeLoRA: Enabling Training-Free LoRA Fusion for Autoregressive Multi-Subject Personalization | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Subject-driven image generation plays a crucial role in applications such as virtual try-on and poster design. Existing approaches typically fine-tune pretrained generative models or apply LoRA-based adaptations for individual subjects. However, these methods struggle with multi-subject personalization, as combining independently adapted modules often requires complex re-tuning or joint optimization. We present FreeLoRA, a simple and generalizable framework that enables training-free fusion of subject-specific LoRA modules for multi-subject personalization. Each LoRA module is adapted on a few images of a specific subject using a Full Token Tuning strategy, where it is applied across all tokens in the prompt to encourage weakly supervised token-content alignment. At inference, we adopt Subject-Aware Inference, activating each module only on its corresponding subject tokens. This enables training-free fusion of multiple personalized subjects within a single image, while mitigating overfitting and mutual interference between subjects. Extensive experiments show that FreeLoRA achieves strong performance in both subject fidelity and prompt consistency. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:16:59 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zheng",
"Peng",
""
],
[
"Wang",
"Ye",
""
],
[
"Ma",
"Rui",
""
],
[
"Wu",
"Zuxuan",
""
]
] |
2507.01793 | Daniel Julian | Daniel Julian and Jes\'us P\'erez-R\'ios | Machine learning prediction of a chemical reaction over 8 decades of energy | 6 pages, 4 figures | null | null | null | physics.chem-ph physics.atom-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent progress in machine learning has sparked increased interest in utilizing this technology to predict the outcomes of chemical reactions. The ultimate aim of such endeavors is to develop a universal model that can predict products for any chemical reaction given reactants and physical conditions. In pursuit of ever more universal chemical predictors, machine learning models for atom-diatom and diatom-diatom reactions have been developed, yet no such models exist for termolecular reactions. Accordingly, we introduce neural networks trained to predict opacity functions of atom recombination reactions. Our models predict the recombination of Sr$^+$ + Cs + Cs $\rightarrow$ SrCs$^+$ + Cs and Sr$^+$ + Cs + Cs $\rightarrow$ Cs$_2$ + Sr$^+$ over multiple orders of magnitude of energy, yielding overall results with a relative error $\lesssim 10\%$. Even far beyond the range of energies seen during training, our models predict the atom recombination reaction rate accurately. As a result, the machine is capable of learning the physics behind the atom recombination reaction dynamics. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:17:17 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Julian",
"Daniel",
""
],
[
"Pérez-Ríos",
"Jesús",
""
]
] |
2507.01794 | Carlo Alberto Barbano | Carlo Alberto Barbano, Benoit Dufumier, Edouard Duchesnay, Marco Grangetto, Pietro Gori | Robust brain age estimation from structural MRI with contrastive learning | 11 pages | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | Estimating brain age from structural MRI has emerged as a powerful tool for characterizing normative and pathological aging. In this work, we explore contrastive learning as a scalable and robust alternative to supervised approaches for brain age estimation. We introduce a novel contrastive loss function, $\mathcal{L}^{exp}$, and evaluate it across multiple public neuroimaging datasets comprising over 20,000 scans. Our experiments reveal four key findings. First, scaling pre-training on diverse, multi-site data consistently improves generalization performance, cutting external mean absolute error (MAE) nearly in half. Second, $\mathcal{L}^{exp}$ is robust to site-related confounds, maintaining low scanner-predictability as training size increases. Third, contrastive models reliably capture accelerated aging in patients with cognitive impairment and Alzheimer's disease, as shown through brain age gap analysis, ROC curves, and longitudinal trends. Lastly, unlike supervised baselines, $\mathcal{L}^{exp}$ maintains a strong correlation between brain age accuracy and downstream diagnostic performance, supporting its potential as a foundation model for neuroimaging. These results position contrastive learning as a promising direction for building generalizable and clinically meaningful brain representations. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:18:03 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Barbano",
"Carlo Alberto",
""
],
[
"Dufumier",
"Benoit",
""
],
[
"Duchesnay",
"Edouard",
""
],
[
"Grangetto",
"Marco",
""
],
[
"Gori",
"Pietro",
""
]
] |
2507.01795 | Lu Zhang | Lizuo Liu, Lu Zhang, Anne Gelb | Neural Entropy-stable conservative flux form neural networks for learning hyperbolic conservation laws | null | null | null | null | math.NA cs.LG cs.NA math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a neural entropy-stable conservative flux form neural network (NESCFN) for learning hyperbolic conservation laws and their associated entropy functions directly from solution trajectories, without requiring any predefined numerical discretization. While recent neural network architectures have successfully integrated classical numerical principles into learned models, most rely on prior knowledge of the governing equations or assume a fixed discretization. Our approach removes this dependency by embedding entropy-stable design principles into the learning process itself, enabling the discovery of physically consistent dynamics in a fully data-driven setting. By jointly learning both the numerical flux function and a corresponding entropy, the proposed method ensures conservation and entropy dissipation, critical for long-term stability and fidelity in the system of hyperbolic conservation laws. Numerical results demonstrate that the method achieves stability and conservation over extended time horizons and accurately captures shock propagation speeds, even without oracle access to future-time solution profiles in the training data. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:18:04 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Liu",
"Lizuo",
""
],
[
"Zhang",
"Lu",
""
],
[
"Gelb",
"Anne",
""
]
] |
2507.01796 | Quentin Remy | Reza Rouzegar (1), Oliver Franke (1), Gal Lemut (1), Oliver Gueckstock (1), Junwei Tong (1), Dieter Engel (2), Xianmin Zhang (3), Georg Woltersdorf (4), Piet Brouwer (1), Tobias Kampfrath (1), Quentin Remy (1) ((1) Department of Physics, Freie Universit\"at Berlin, Berlin, Germany, (2) Max-Born-Institut f\"ur nichtlineare Optik und Kurzzeitspektroskopie, Berlin, Germany, (3) Key Laboratory for Anisotropy and Texture of Materials (Ministry of Education), School of Material Science and Engineering, Northeastern University, Shenyang, China, (4) Institut f\"ur Physik, Martin-Luther-Universit\"at Halle, Halle, Germany) | Femtosecond signatures of optically induced magnons before ultrafast demagnetization | null | null | null | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | Optically induced demagnetization of 3d metallic ferromagnets proceeds as fast as ~100 fs and is a crucial prerequisite for spintronic applications, such as ultrafast magnetization switching and spin transport. On the 100 fs time scale, the magnetization dynamics is widely understood in the context of temperature models considering energy transfers between conduction electrons, magnons and crystal lattice. However, on even faster time scales, the flow of both angular momentum and energy between these subsystems has so far not been studied. Here, we measure ultrafast demagnetization by ultrabroadband THz-emission spectroscopy. We find that the rate of change of the magnetization does not rise instantaneously, but on a time scale as short as 10 fs. This rise is a signature that a transfer of angular momentum from the magnons to conduction electrons proceeds in less than 10 fs, before substantial demagnetization has happened. We further conclude that most of the spin dissipated by the lattice is transferred via magnon-lattice rather than electron-lattice interaction. These results show that the limiting speed of magnetization dynamics is not demagnetization, as generally believed, and harnessing the earliest magnon dynamics could be a new route towards an even faster spintronics. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:18:46 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Rouzegar",
"Reza",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Franke",
"Oliver",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Lemut",
"Gal",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Gueckstock",
"Oliver",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Tong",
"Junwei",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Engel",
"Dieter",
"",
"Max-Born-Institut für nichtlineare Optik und Kurzzeitspektroskopie, Berlin, Germany"
],
[
"Zhang",
"Xianmin",
"",
"Key Laboratory for Anisotropy and Texture of Materials"
],
[
"Woltersdorf",
"Georg",
"",
"Institut für Physik, Martin-Luther-Universität Halle, Halle, Germany"
],
[
"Brouwer",
"Piet",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Kampfrath",
"Tobias",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
],
[
"Remy",
"Quentin",
"",
"Department of Physics, Freie Universität Berlin, Berlin, Germany"
]
] |
2507.01797 | Jefferson Baudin | Jefferson Baudin | On the Euler characteristic of weakly ordinary varieties of maximal Albanese dimension | Comments welcome! | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that a smooth proper weakly ordinary variety $X$ of maximal Albanese dimension satisfies $\chi(X, \omega_X) \geq 0$. We also show that if $X$ is not of general type, then $\chi(X, \omega_X) = 0$ and the Albanese image of $X$ is fibered by abelian varieties. The proof uses the positive characteristic generic vanishing theory developed by Hacon-Patakfalvi, as well as our recent Witt vector version of Grauert-Riemenschneider vanishing. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:19:23 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Baudin",
"Jefferson",
""
]
] |
2507.01798 | Valeri Makarov | Valeri V. Makarov | Kinematic Distortions of the High-Redshift Universe as Seen from Quasar Proper Motions | Author's version of the paper published in Nature Astronomy on July 3, 2025 much extended with additional information on validation/verification, methodology, technical implementation | null | 10.1038/s41550-025-02591-x | null | astro-ph.GA astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Advances in optical astrometry allow us to infer the non-radial kinematic structure of the Universe directly from observations. Here I use a supervised machine learning neural network method to predict 1.57 million redshifts based on several photometric and metadata classifier parameters from the unWISE mid-infrared database and from Gaia. These estimates are used to divide the sample into three redshift bins: 1-2, 2-3, and $>3$. For each subset, all available Gaia proper motions are used in a global vector spherical harmonic solution to degree 3 (30 fitting vector functions). I find significant differences in a few fitted proper motion patterns at different redshifts. The largest signals are seen in the comparison of the vector spherical harmonic fits for the 1-2 and 2-3 redshift bins. The significant harmonics include a rigid spin, a dipole glide from the north Galactic pole to the south and an additional quadrupole distortion. Validation tests with filtered subsamples indicate that the detected effect can be caused by hidden systematic errors in astrometry. The results are verified by using an independent source of redshifts and computing the observer's Galactocentric acceleration. This study offers a new observational test of alternative cosmological models. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:19:46 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Makarov",
"Valeri V.",
""
]
] |
2507.01799 | Steffen Schieler | Steffen Schieler, Sebastian Semper, Christian Schneider, Reiner Thom\"a | Measurement-based Evaluation of CNN-based Detection and Estimation for ISAC Systems | 2025 IEEE International Radar Conference (RADAR) | null | 10.1109/RADAR52380.2025.11031730 | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In wireless sensing applications, such as ISAC, one of the first crucial signal processing steps is the detection and estimation targets from a channel estimate. Effective algorithms in this context must be robust across a broad SNR range, capable of handling an unknown number of targets, and computationally efficient for real-time implementation. During the last decade, different Machine Learning methods have emerged as promising solutions, either as standalone models or as complementing existing techniques. However, since models are often trained and evaluated on synthetic data from existing models, applying them to measurement is challenging. All the while, training directly on measurement data is prohibitive in complex propagation scenarios as a groundtruth is not available. Therefore, in this paper, we train a CNN approach for target detection and estimation on synthetic data and evaluate it on measurement data from a suburban outdoor measurement. Using knowledge of the environment as well as available groundtruth positions, we study the detection probability and accuracy of our approach. The results demonstrate that our approach works on measurement data and is suitable for joint detection and estimation of sensing targets in ISAC systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:19:52 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Schieler",
"Steffen",
""
],
[
"Semper",
"Sebastian",
""
],
[
"Schneider",
"Christian",
""
],
[
"Thomä",
"Reiner",
""
]
] |
2507.01800 | Shengli Zhou | Shengli Zhou, Jianuo Zhu, Qilin Huang, Fangjing Wang, Yanfu Zhang, Feng Zheng | HCNQA: Enhancing 3D VQA with Hierarchical Concentration Narrowing Supervision | ICANN 2025 | null | null | null | cs.CV cs.MM | http://creativecommons.org/licenses/by/4.0/ | 3D Visual Question-Answering (3D VQA) is pivotal for models to perceive the physical world and perform spatial reasoning. Answer-centric supervision is a commonly used training method for 3D VQA models. Many models that utilize this strategy have achieved promising results in 3D VQA tasks. However, the answer-centric approach only supervises the final output of models and allows models to develop reasoning pathways freely. The absence of supervision on the reasoning pathway enables the potential for developing superficial shortcuts through common patterns in question-answer pairs. Moreover, although slow-thinking methods advance large language models, they suffer from underthinking. To address these issues, we propose \textbf{HCNQA}, a 3D VQA model leveraging a hierarchical concentration narrowing supervision method. By mimicking the human process of gradually focusing from a broad area to specific objects while searching for answers, our method guides the model to perform three phases of concentration narrowing through hierarchical supervision. By supervising key checkpoints on a general reasoning pathway, our method can ensure the development of a rational and effective reasoning pathway. Extensive experimental results demonstrate that our method can effectively ensure that the model develops a rational reasoning pathway and performs better. The code is available at https://github.com/JianuoZhu/HCNQA. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:20:08 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Zhou",
"Shengli",
""
],
[
"Zhu",
"Jianuo",
""
],
[
"Huang",
"Qilin",
""
],
[
"Wang",
"Fangjing",
""
],
[
"Zhang",
"Yanfu",
""
],
[
"Zheng",
"Feng",
""
]
] |
2507.01801 | Bin Rao | Bin Rao, Haicheng Liao, Yanchen Guan, Chengyue Wang, Bonan Wang, Jiaxun Zhang, and Zhenning Li | AMD: Adaptive Momentum and Decoupled Contrastive Learning Framework for Robust Long-Tail Trajectory Prediction | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately predicting the future trajectories of traffic agents is essential in autonomous driving. However, due to the inherent imbalance in trajectory distributions, tail data in natural datasets often represents more complex and hazardous scenarios. Existing studies typically rely solely on a base model's prediction error, without considering the diversity and uncertainty of long-tail trajectory patterns. We propose an adaptive momentum and decoupled contrastive learning framework (AMD), which integrates unsupervised and supervised contrastive learning strategies. By leveraging an improved momentum contrast learning (MoCo-DT) and decoupled contrastive learning (DCL) module, our framework enhances the model's ability to recognize rare and complex trajectories. Additionally, we design four types of trajectory random augmentation methods and introduce an online iterative clustering strategy, allowing the model to dynamically update pseudo-labels and better adapt to the distributional shifts in long-tail data. We propose three different criteria to define long-tail trajectories and conduct extensive comparative experiments on the nuScenes and ETH$/$UCY datasets. The results show that AMD not only achieves optimal performance in long-tail trajectory prediction but also demonstrates outstanding overall prediction accuracy. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:20:40 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Rao",
"Bin",
""
],
[
"Liao",
"Haicheng",
""
],
[
"Guan",
"Yanchen",
""
],
[
"Wang",
"Chengyue",
""
],
[
"Wang",
"Bonan",
""
],
[
"Zhang",
"Jiaxun",
""
],
[
"Li",
"Zhenning",
""
]
] |
2507.01802 | Katharina Beckh | Katharina Beckh, Elisa Studeny, Sujan Sai Gannamaneni, Dario Antweiler, Stefan R\"uping | The Anatomy of Evidence: An Investigation Into Explainable ICD Coding | Accepted to ACL 2025 Findings | null | null | null | cs.CL cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Automatic medical coding has the potential to ease documentation and billing processes. For this task, transparency plays an important role for medical coders and regulatory bodies, which can be achieved using explainability methods. However, the evaluation of these approaches has been mostly limited to short text and binary settings due to a scarcity of annotated data. Recent efforts by Cheng et al. (2023) have introduced the MDACE dataset, which provides a valuable resource containing code evidence in clinical records. In this work, we conduct an in-depth analysis of the MDACE dataset and perform plausibility evaluation of current explainable medical coding systems from an applied perspective. With this, we contribute to a deeper understanding of automatic medical coding and evidence extraction. Our findings reveal that ground truth evidence aligns with code descriptions to a certain degree. An investigation into state-of-the-art approaches shows a high overlap with ground truth evidence. We propose match measures and highlight success and failure cases. Based on our findings, we provide recommendations for developing and evaluating explainable medical coding systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:21:29 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Beckh",
"Katharina",
""
],
[
"Studeny",
"Elisa",
""
],
[
"Gannamaneni",
"Sujan Sai",
""
],
[
"Antweiler",
"Dario",
""
],
[
"Rüping",
"Stefan",
""
]
] |
2507.01803 | Leyang Xue | Leyang Xue, Meghana Madhyastha, Randal Burns, Myungjin Lee, Mahesh K. Marina | Towards Decentralized and Sustainable Foundation Model Training with the Edge | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Foundation models are at the forefront of AI research, appealing for their ability to learn from vast datasets and cater to diverse tasks. Yet, their significant computational demands raise issues of environmental impact and the risk of centralized control in their development. We put forward a vision towards decentralized and sustainable foundation model training that leverages the collective compute of sparingly used connected edge AI devices. We present the rationale behind our vision, particularly in support of its sustainability benefit. We further outline a set of challenges that need to be addressed to turn this vision into reality. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:21:40 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Xue",
"Leyang",
""
],
[
"Madhyastha",
"Meghana",
""
],
[
"Burns",
"Randal",
""
],
[
"Lee",
"Myungjin",
""
],
[
"Marina",
"Mahesh K.",
""
]
] |
2507.01804 | Richard Tol | Richard S.J. Tol | Meta-emulation: An application to the social cost of carbon | null | null | null | null | econ.EM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A large database of published model results is used to estimate the distribution of the social cost of carbon as a function of the underlying assumptions. The literature on the social cost of carbon deviates in its assumptions from the literatures on the impacts of climate change, discounting, and risk aversion. The proposed meta-emulator corrects this. The social cost of carbon is higher than reported in the literature. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:23:54 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Tol",
"Richard S. J.",
""
]
] |
2507.01805 | Alejandro Sosa Welford | Alejandro Sosa Welford, Leonardo Pepino | A Dataset for Automatic Assessment of TTS Quality in Spanish | 5 pages, 2 figures. Accepted at Interspeech 2025 | null | null | null | cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | This work addresses the development of a database for the automatic assessment of text-to-speech (TTS) systems in Spanish, aiming to improve the accuracy of naturalness prediction models. The dataset consists of 4,326 audio samples from 52 different TTS systems and human voices and is, up to our knowledge, the first of its kind in Spanish. To label the audios, a subjective test was designed based on the ITU-T Rec. P.807 standard and completed by 92 participants. Furthermore, the utility of the collected dataset was validated by training automatic naturalness prediction systems. We explored two approaches: fine-tuning an existing model originally trained for English, and training small downstream networks on top of frozen self-supervised speech models. Our models achieve a mean absolute error of 0.8 on a five-point MOS scale. Further analysis demonstrates the quality and diversity of the developed dataset, and its potential to advance TTS research in Spanish. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:24:47 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Welford",
"Alejandro Sosa",
""
],
[
"Pepino",
"Leonardo",
""
]
] |
2507.01806 | Reza Arabpour | Reza Arabpour, Haitz S\'aez de Oc\'ariz Borde, Anastasis Kratsios | LoRA Fine-Tuning Without GPUs: A CPU-Efficient Meta-Generation Framework for LLMs | 5-page main paper (excluding references) + 11-page appendix, 3 tables, 1 figure. Accepted to ICML 2025 Workshop on Efficient Systems for Foundation Models | null | null | null | cs.LG cs.AI cs.CL stat.ML | http://creativecommons.org/licenses/by/4.0/ | Low-Rank Adapters (LoRAs) have transformed the fine-tuning of Large Language Models (LLMs) by enabling parameter-efficient updates. However, their widespread adoption remains limited by the reliance on GPU-based training. In this work, we propose a theoretically grounded approach to LoRA fine-tuning designed specifically for users with limited computational resources, particularly those restricted to standard laptop CPUs. Our method learns a meta-operator that maps any input dataset, represented as a probability distribution, to a set of LoRA weights by leveraging a large bank of pre-trained adapters for the Mistral-7B-Instruct-v0.2 model. Instead of performing new gradient-based updates, our pipeline constructs adapters via lightweight combinations of existing LoRAs directly on CPU. While the resulting adapters do not match the performance of GPU-trained counterparts, they consistently outperform the base Mistral model on downstream tasks, offering a practical and accessible alternative to traditional GPU-based fine-tuning. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:24:47 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Arabpour",
"Reza",
""
],
[
"Borde",
"Haitz Sáez de Ocáriz",
""
],
[
"Kratsios",
"Anastasis",
""
]
] |
2507.01807 | Salom\'e Grouffal | S. Grouffal, A. Santerne, V. Bourrier, V. Kunovac, C. Dressing, B. Akinsanmi, C. Armstrong, S. Baliwal, O. Balsalobre-Ruza, S.C.C. Barros, D. Bayliss, I.J.M. Crossfield, O. Demangeon, X. Dumusque, S. Giacalone, C.K. Harada, H. Isaacson, H. Kellermann, J. Lillo-Box, J.Llama, A. Mortier, E. Palle, A.S. Rajpurohit, M. Rice, N.C. Santos, J.V. Seidel, R. Sharma, S.G. Sousa, L. Thomas, E.V. Turtelboom, S. Udry, and P.J. Wheatley | The star HIP 41378 potentially misaligned with its cohort of long-period planets | 15 pages, 9 figures. Submitted to A&A | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | The obliquity between the stellar spin axis and the planetary orbit, detected via the Rossiter-McLaughlin (RM) effect, is a tracer of the formation history of planetary systems. While obliquity measurements have been extensively applied to hot Jupiters and short-period planets, they remain rare for cold and long-period planets due to observational challenges, particularly their long transit durations. We report the detection of the RM effect for the 19-hour-long transit of HIP 41378 f, a temperate giant planet on a 542-day orbit, observed through a worldwide spectroscopic campaign. We measure a slight projected obliquity of 21 $\pm$ 8 degrees and a significant 3D spin-orbit angle of 52 $\pm$ 6 degrees, based on the measurement of the stellar rotation period. HIP 41378 f is part of a 5-transiting planetary system with planets close to mean motion resonances. The observed misalignment likely reflects a primordial tilt of the stellar spin axis relative to the protoplanetary disk, rather than dynamical interactions. HIP 41378 f is the first non-eccentric long-period (P>100 days) planet observed with the RM effect, opening new constraints on planetary formation theories. This observation should motivate the exploration of planetary obliquities across a longer range of orbital distances through international collaboration. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:24:48 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Grouffal",
"S.",
""
],
[
"Santerne",
"A.",
""
],
[
"Bourrier",
"V.",
""
],
[
"Kunovac",
"V.",
""
],
[
"Dressing",
"C.",
""
],
[
"Akinsanmi",
"B.",
""
],
[
"Armstrong",
"C.",
""
],
[
"Baliwal",
"S.",
""
],
[
"Balsalobre-Ruza",
"O.",
""
],
[
"Barros",
"S. C. C.",
""
],
[
"Bayliss",
"D.",
""
],
[
"Crossfield",
"I. J. M.",
""
],
[
"Demangeon",
"O.",
""
],
[
"Dumusque",
"X.",
""
],
[
"Giacalone",
"S.",
""
],
[
"Harada",
"C. K.",
""
],
[
"Isaacson",
"H.",
""
],
[
"Kellermann",
"H.",
""
],
[
"Lillo-Box",
"J.",
""
],
[
"Llama",
"J.",
""
],
[
"Mortier",
"A.",
""
],
[
"Palle",
"E.",
""
],
[
"Rajpurohit",
"A. S.",
""
],
[
"Rice",
"M.",
""
],
[
"Santos",
"N. C.",
""
],
[
"Seidel",
"J. V.",
""
],
[
"Sharma",
"R.",
""
],
[
"Sousa",
"S. G.",
""
],
[
"Thomas",
"L.",
""
],
[
"Turtelboom",
"E. V.",
""
],
[
"Udry",
"S.",
""
],
[
"Wheatley",
"P. J.",
""
]
] |
2507.01808 | Joshua Shank | Xiaoyu Ji, Jessica Shorland, Joshua Shank, Pascal Delpe-Brice, Latanya Sweeney, Jan Allebach, and Ali Shakouri | Empowering Manufacturers with Privacy-Preserving AI Tools: A Case Study in Privacy-Preserving Machine Learning to Solve Real-World Problems | 20 pages, 11 figures, 30 references | null | null | null | cs.CR cs.AI cs.CV cs.ET | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Small- and medium-sized manufacturers need innovative data tools but, because of competition and privacy concerns, often do not want to share their proprietary data with researchers who might be interested in helping. This paper introduces a privacy-preserving platform by which manufacturers may safely share their data with researchers through secure methods, so that those researchers then create innovative tools to solve the manufacturers' real-world problems, and then provide tools that execute solutions back onto the platform for others to use with privacy and confidentiality guarantees. We illustrate this problem through a particular use case which addresses an important problem in the large-scale manufacturing of food crystals, which is that quality control relies on image analysis tools. Previous to our research, food crystals in the images were manually counted, which required substantial and time-consuming human efforts, but we have developed and deployed a crystal analysis tool which makes this process both more rapid and accurate. The tool enables automatic characterization of the crystal size distribution and numbers from microscope images while the natural imperfections from the sample preparation are automatically removed; a machine learning model to count high resolution translucent crystals and agglomeration of crystals was also developed to aid in these efforts. The resulting algorithm was then packaged for real-world use on the factory floor via a web-based app secured through the originating privacy-preserving platform, allowing manufacturers to use it while keeping their proprietary data secure. After demonstrating this full process, future directions are also explored. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:25:43 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Ji",
"Xiaoyu",
""
],
[
"Shorland",
"Jessica",
""
],
[
"Shank",
"Joshua",
""
],
[
"Delpe-Brice",
"Pascal",
""
],
[
"Sweeney",
"Latanya",
""
],
[
"Allebach",
"Jan",
""
],
[
"Shakouri",
"Ali",
""
]
] |
2507.01809 | Ma Guancong | Xiao-Ming Wang, Jiaying Xu, Xulong Wang, Zhen Li, Guancong Ma | Topological Braiding of Bloch Eigenmodes Protected by Non-Abelian Quaternion Invariants | null | null | null | null | quant-ph physics.app-ph physics.class-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Braiding has attracted significant attention in physics because of its important role in describing the fundamental exchange of particles. Infusing the braiding with topological protection will make it robust against imperfections and perturbations, but such topological braiding is believed to be possible only in interacting quantum systems, e.g., topological superconductors. Here, we propose and demonstrate a new strategy of topological braiding that emerges from non-Abelian topological insulators, a class of recently discovered multi-band topological phase. We unveil a mathematical connection between braiding and non-Abelian quaternion invariants, by which Bloch eigenmodes under parallel transport produce braid sequences protected by the non-Abelian band topology. The braiding is also associated with geometric phases quantized over half the Brillouin zone. This new type of non-Abelian topological braiding is experimentally realized in acoustic systems with periodic synthetic dimensions. The results show that the principle discovered here is a new strategy towards topological braiding and can be extended for other types of classical waves and non-interacting quantum systems. | [
{
"version": "v1",
"created": "Wed, 02 Jul 2025 15:26:19 GMT"
}
] | 2025-07-03T00:00:00 | [
[
"Wang",
"Xiao-Ming",
""
],
[
"Xu",
"Jiaying",
""
],
[
"Wang",
"Xulong",
""
],
[
"Li",
"Zhen",
""
],
[
"Ma",
"Guancong",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.