id
stringlengths 9
16
| submitter
stringlengths 1
64
⌀ | authors
stringlengths 3
62.8k
| title
stringlengths 1
382
| comments
stringlengths 1
1.15k
⌀ | journal-ref
stringlengths 1
557
⌀ | doi
stringlengths 7
153
⌀ | report-no
stringlengths 1
509
⌀ | categories
stringlengths 5
125
| license
stringclasses 9
values | abstract
stringlengths 6
6.09k
| versions
listlengths 1
187
| update_date
timestamp[s]date 2007-05-23 00:00:00
2025-04-21 00:00:00
| authors_parsed
sequencelengths 1
2.83k
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2504.11936 | Xia Deng | Xia Deng, Shen Chen, Jiale Zhou, Lei Li | Mind2Matter: Creating 3D Models from EEG Signals | null | null | null | null | cs.GR cs.HC eess.SP | http://creativecommons.org/licenses/by/4.0/ | The reconstruction of 3D objects from brain signals has gained significant
attention in brain-computer interface (BCI) research. Current research
predominantly utilizes functional magnetic resonance imaging (fMRI) for 3D
reconstruction tasks due to its excellent spatial resolution. Nevertheless, the
clinical utility of fMRI is limited by its prohibitive costs and inability to
support real-time operations. In comparison, electroencephalography (EEG)
presents distinct advantages as an affordable, non-invasive, and mobile
solution for real-time brain-computer interaction systems. While recent
advances in deep learning have enabled remarkable progress in image generation
from neural data, decoding EEG signals into structured 3D representations
remains largely unexplored. In this paper, we propose a novel framework that
translates EEG recordings into 3D object reconstructions by leveraging neural
decoding techniques and generative models. Our approach involves training an
EEG encoder to extract spatiotemporal visual features, fine-tuning a large
language model to interpret these features into descriptive multimodal outputs,
and leveraging generative 3D Gaussians with layout-guided control to synthesize
the final 3D structures. Experiments demonstrate that our model captures
salient geometric and semantic features, paving the way for applications in
brain-computer interfaces (BCIs), virtual reality, and neuroprosthetics. Our
code is available in https://github.com/sddwwww/Mind2Matter.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 10:16:03 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 12:17:59 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Deng",
"Xia",
""
],
[
"Chen",
"Shen",
""
],
[
"Zhou",
"Jiale",
""
],
[
"Li",
"Lei",
""
]
] |
2504.12283 | Wenqi Ke | Ignatios Antoniadis, John Ellis, Wenqi Ke, Dimitri V. Nanopoulos,
Keith A. Olive | How Accidental was Inflation? | 27 pages, 17 figures | null | null | UMN-TH-4418/25, FTPI-MINN-25/03, KCL-PH-TH/2025-09, CERN-TH-2025-076 | hep-ph astro-ph.CO gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | Data on the cosmic microwave background (CMB) are discriminating between
different models of inflation, disfavoring simple monomial potentials whilst
being consistent with models whose predictions resemble those of the
Starobinsky $R + R^2$ cosmological model. However, this model may suffer from
theoretical problems, since it requires a large initial field value,
threatening the validity of the effective field theory. This is quantified by
the Swampland Distance Conjecture, which predicts the appearance of a tower of
light states associated with an effective ultra-violet cutoff. This could be
lower than the inflation scale for cases with an extended period of inflation,
leading to an additional problem of initial conditions. No-scale supergravity
models can reproduce the predictions of the Starobinsky model and accommodate
the CMB data at the expense of fine-tuning of parameters at the level of
$10^{-5}$. Here, we propose a solution to this problem based on an explicit
realisation of the Starobinsky model in string theory, where this `deformation'
parameter is calculable and takes a value of order of the one corresponding to
the Starobinsky inflaton potential. Within this range, there are parameter
values that accommodate more easily the combination of Planck, ACT and DESI BAO
data, while also restricting the range of possible inflaton field values,
thereby avoiding the swampland problem and predicting that the initial
conditions for inflation compatible with the CMB data are generic.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 17:45:53 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Antoniadis",
"Ignatios",
""
],
[
"Ellis",
"John",
""
],
[
"Ke",
"Wenqi",
""
],
[
"Nanopoulos",
"Dimitri V.",
""
],
[
"Olive",
"Keith A.",
""
]
] |
2504.12309 | Yi-De Lin | Yi-De Lin, Guan-Ze Liao | Large Language Model-Based Knowledge Graph System Construction for
Sustainable Development Goals: An AI-Based Speculative Design Perspective | This is a minor revision: fixed a typo in the abstract (time range)
and corrected minor textual errors | null | null | null | cs.CY cs.AI cs.CL cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | From 2000 to 2015, the UN's Millennium Development Goals guided global
priorities. The subsequent Sustainable Development Goals (SDGs) adopted a more
dynamic approach, with annual indicator updates. As 2030 nears and progress
lags, innovative acceleration strategies are critical. This study develops an
AI-powered knowledge graph system to analyze SDG interconnections, discover
potential new goals, and visualize them online. Using official SDG texts,
Elsevier's keyword dataset, and 1,127 TED Talk transcripts (2020.01-2024.04), a
pilot on 269 talks from 2023 applies AI-speculative design, large language
models, and retrieval-augmented generation. Key findings include: (1) Heatmap
analysis reveals strong associations between Goal 10 and Goal 16, and minimal
coverage of Goal 6. (2) In the knowledge graph, simulated dialogue over time
reveals new central nodes, showing how richer data supports divergent thinking
and goal clarity. (3) Six potential new goals are proposed, centered on equity,
resilience, and technology-driven inclusion. This speculative-AI framework
offers fresh insights for policymakers and lays groundwork for future
multimodal and cross-system SDG applications.
| [
{
"version": "v1",
"created": "Sat, 5 Apr 2025 17:11:53 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 05:38:41 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Lin",
"Yi-De",
""
],
[
"Liao",
"Guan-Ze",
""
]
] |
2504.12340 | Bumned Soodchomshom Dr. | Bumned Soodchomshom | Particle-Hole Creation in Condensed Matter: A Conceptual Framework for
Modeling Money-Debt Dynamics in Economics | 12 pages,1 figure, 2 table, section 4.5 added | null | null | null | econ.GN q-fin.EC quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a field-theoretic framework that models money-debt dynamics in
economic systems through a direct analogy to particle-hole creation in
condensed matter physics. In this formulation, issuing credit generates a
symmetric pair-money as a particle-like excitation and debt as its hole-like
counterpart-embedded within a monetary vacuum field. The model is formalized
via a second-quantized Hamiltonian that incorporates time-dependent
perturbations to represent real-world effects such as interest and profit,
which drive asymmetry and systemic imbalance. This framework successfully
captures both macroeconomic phenomena, including quantitative easing (QE) and
gold-backed monetary regimes, and microeconomic credit creation, under a
unified quantum-like formalism. In particular, QE is interpreted as generating
entangled-like pairs of currency and bonds, exhibiting systemic correlations
akin to nonlocal quantum interactions. Asset-backed systems, on the other hand,
are modeled as coherent superpositions that collapse upon use. This approach
provides physicists with a rigorous and intuitive toolset to analyze economic
behavior using many-body theory, laying the groundwork for a new class of
models in econophysics and interdisciplinary field analysis.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 05:37:15 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 04:07:22 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Soodchomshom",
"Bumned",
""
]
] |
2504.12354 | Vinay Shukla | Vinay Shukla, Prachee Sharma, Ryan Rossi, Sungchul Kim, Tong Yu,
Aditya Grover | WaterFlow: Learning Fast & Robust Watermarks using Stable Diffusion | null | null | null | null | eess.IV cs.AI | http://creativecommons.org/licenses/by/4.0/ | The ability to embed watermarks in images is a fundamental problem of
interest for computer vision, and is exacerbated by the rapid rise of generated
imagery in recent times. Current state-of-the-art techniques suffer from
computational and statistical challenges such as the slow execution speed for
practical deployments. In addition, other works trade off fast watermarking
speeds but suffer greatly in their robustness or perceptual quality. In this
work, we propose WaterFlow (WF), a fast and extremely robust approach for high
fidelity visual watermarking based on a learned latent-dependent watermark. Our
approach utilizes a pretrained latent diffusion model to encode an arbitrary
image into a latent space and produces a learned watermark that is then planted
into the Fourier Domain of the latent. The transformation is specified via
invertible flow layers that enhance the expressivity of the latent space of the
pre-trained model to better preserve image quality while permitting robust and
tractable detection. Most notably, WaterFlow demonstrates state-of-the-art
performance on general robustness and is the first method capable of
effectively defending against difficult combination attacks. We validate our
findings on three widely used real and generated datasets: MS-COCO,
DiffusionDB, and WikiArt.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 23:27:52 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 02:12:50 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Shukla",
"Vinay",
""
],
[
"Sharma",
"Prachee",
""
],
[
"Rossi",
"Ryan",
""
],
[
"Kim",
"Sungchul",
""
],
[
"Yu",
"Tong",
""
],
[
"Grover",
"Aditya",
""
]
] |
2504.12403 | Alexander Marchuk | Vladimir P. Reshetnikov, Ilia V. Chugunov, Alexander A. Marchuk,
Aleksandr V. Mosenkov, Matvey D. Kozlov, Sergey S. Savchenko, Dmitry I.
Makarov, Aleksandra V. Antipova, Anastasia M. Sypkova | Galactic warps: from cosmic noon to the current epoch | Accepted for publication in A&A Letters | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Approximately half of all disc galaxies exhibit appreciable warps in both
their stellar and HI discs. The typical warp amplitude is small (a few
degrees), and only becomes noticeable at the periphery of the galaxy disc. As a
result, warps remain a relatively poorly studied phenomenon. In this study, we
investigate a large sample of distant edge-on galaxies (approximately 1,000
objects) in order to examine the frequency and characteristics of stellar disc
warps up to a redshift of $z \sim 2$. For the selected galaxies, we utilize HST
data from the Cosmic Evolution Survey field and JWST observations from the
Cosmic Dawn Center Archive. We measured the properties of disc warps and
investigate their evolution as a function of redshift. Our results indicate a
potential evolution in the observed frequency of strong S-shaped warps (with an
amplitude greater than 4$^\circ$) in stellar discs as a function of redshift.
At $z \approx 2$, the frequency of strong warps reaches approximately 50%,
while at $z \approx 0$, this fraction decreases to around 10-15%. We attribute
the observed evolution in the occurrence of strong warps to the changing
frequency of galaxy interactions and mergers. If galaxy interactions represent
one of the primary mechanisms responsible for the formation of warps, then the
prevalence of vertical disc deformations should increase in tandem with the
rising interaction and merger rate.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 18:07:00 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 10:44:46 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Reshetnikov",
"Vladimir P.",
""
],
[
"Chugunov",
"Ilia V.",
""
],
[
"Marchuk",
"Alexander A.",
""
],
[
"Mosenkov",
"Aleksandr V.",
""
],
[
"Kozlov",
"Matvey D.",
""
],
[
"Savchenko",
"Sergey S.",
""
],
[
"Makarov",
"Dmitry I.",
""
],
[
"Antipova",
"Aleksandra V.",
""
],
[
"Sypkova",
"Anastasia M.",
""
]
] |
2504.12432 | You Chang | You Chang, Jose L. Gonzales, Mossa Merhi Reimert, Erik Rattenborg,
Mart C.M. de Jong, Beate Conrady | Assessing the Spatial and Temporal Risk of HPAIV Transmission to Danish
Cattle via Wild Birds | 12 pages, 5 figures | null | null | null | q-bio.PE q-bio.QM | http://creativecommons.org/licenses/by/4.0/ | A highly pathogenic avian influenza (HPAI) panzootic has severely impacted
wild bird populations worldwide, with documented (zoonotic) transmission to
mammals, including humans. Ongoing HPAI outbreaks on U.S. cattle farms have
raised concerns about potential spillover of virus from birds to cattle in
other countries, including Denmark. In the EU, the Bird Flu Radar tool,
coordinated by EFSA, monitors the spatio-temporal risk of HPAIV infection in
wild bird populations. A preparedness tool to assess the spillover risk to the
cattle industry is currently lacking, despite its critical importance. This
study aims to assess the temporal and spatial risk of HPAI virus (HPAIV)
spillover from wild birds, particularly waterfowl, into cattle populations in
Denmark. To support this assessment, a spillover transmission model is
developed by integrating two well-established surveillance tools, eBird and
Bird Flu Radar, in combination with global cattle density data. The generated
quantitative risk maps reveal the heterogeneous temporal and spatial
distribution of HPAIV spillover risk from wild birds to cattle across Denmark.
The highest risk periods are observed during calendar weeks 50 to 10. The
estimated total number of spillover cases nationwide is 1.93 (95% CI: 0.48,
4.98) in 2024, and 0.62 cases (95% CI: 0.15, 1.25) in 2025 (up to April). These
risk estimates provide valuable insights to support veterinary contingency
planning and enable targeted allocation of resources in highrisk areas for the
early detection of HPAIV in cattle.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 19:01:19 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 06:25:44 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Chang",
"You",
""
],
[
"Gonzales",
"Jose L.",
""
],
[
"Reimert",
"Mossa Merhi",
""
],
[
"Rattenborg",
"Erik",
""
],
[
"de Jong",
"Mart C. M.",
""
],
[
"Conrady",
"Beate",
""
]
] |
2504.12435 | Mihoub Bouderbala | Mihoub Bouderbala | On certain sums involving the largest prime factor over integer
sequences | preprints, 7 pages | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given an integer $ n \geq 2 $, its prime factorization is expressed as $ n =
\prod p_i^{a_i} $. We define the function $ f(n) $ as the smallest positive
integer satisfying the following condition: \[
\nu_{p}\left(\frac{f(n)!}{n}\right) \geq 1, \quad \forall p \in \{p_1, p_2,
\dots, p_s\}, \] where $ \nu_{p}(m) $ denotes the $ p $-adic valuation of $ m
$. The main objective of this paper is to derive an asymptotic formula for both
sums $ \sum_{n\leq x} f(n) $ and $ \sum_{n \leq x, n \in S_k} f(n) $, where $
S_k $ denotes the set of all $ k $-free integers.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 19:09:54 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 06:59:52 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Bouderbala",
"Mihoub",
""
]
] |
2504.12463 | Ashwinee Panda | Ashwinee Panda, Vatsal Baherwani, Zain Sarwar, Benjamin Therien,
Supriyo Chakraborty, Tom Goldstein | Dense Backpropagation Improves Training for Sparse Mixture-of-Experts | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Mixture of Experts (MoE) pretraining is more scalable than dense Transformer
pretraining, because MoEs learn to route inputs to a sparse set of their
feedforward parameters. However, this means that MoEs only receive a sparse
backward update, leading to training instability and suboptimal performance. We
present a lightweight approximation method that gives the MoE router a dense
gradient update while continuing to sparsely activate its parameters. Our
method, which we refer to as Default MoE, substitutes missing expert
activations with default outputs consisting of an exponential moving average of
expert outputs previously seen over the course of training. This allows the
router to receive signals from every expert for each token, leading to
significant improvements in training performance. Our Default MoE outperforms
standard TopK routing in a variety of settings without requiring significant
computational overhead. Code: https://github.com/vatsal0/default-moe.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 19:55:36 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 02:58:46 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Panda",
"Ashwinee",
""
],
[
"Baherwani",
"Vatsal",
""
],
[
"Sarwar",
"Zain",
""
],
[
"Therien",
"Benjamin",
""
],
[
"Chakraborty",
"Supriyo",
""
],
[
"Goldstein",
"Tom",
""
]
] |
2504.12643 | Xufeng Huang | Hang Ji, Tao Ni, Xufeng Huang, Tao Luo, Xin Zhan, Junbo Chen | RoPETR: Improving Temporal Camera-Only 3D Detection by Integrating
Enhanced Rotary Position Embedding | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This technical report introduces a targeted improvement to the StreamPETR
framework, specifically aimed at enhancing velocity estimation, a critical
factor influencing the overall NuScenes Detection Score. While StreamPETR
exhibits strong 3D bounding box detection performance as reflected by its high
mean Average Precision our analysis identified velocity estimation as a
substantial bottleneck when evaluated on the NuScenes dataset. To overcome this
limitation, we propose a customized positional embedding strategy tailored to
enhance temporal modeling capabilities. Experimental evaluations conducted on
the NuScenes test set demonstrate that our improved approach achieves a
state-of-the-art NDS of 70.86% using the ViT-L backbone, setting a new
benchmark for camera-only 3D object detection.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 05:05:31 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 06:53:59 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ji",
"Hang",
""
],
[
"Ni",
"Tao",
""
],
[
"Huang",
"Xufeng",
""
],
[
"Luo",
"Tao",
""
],
[
"Zhan",
"Xin",
""
],
[
"Chen",
"Junbo",
""
]
] |
2504.12668 | Boyuan Xu | Boyuan Xu, Liyi Bai, Shenzhen Xu, Qisheng Wu | Observing Nucleation and Crystallization of Rocksalt LiF from Molten
State through Molecular Dynamics Simulations with Refined Machine-Learned
Force Field | null | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Lithium fluoride (LiF) is a critical component for stabilizing lithium metal
anode and high-voltage cathodes towards the next-generation high-energy-density
lithium batteries. Recent modeling study reported the formation of wurtzite LiF
below about 550 K (J. Am. Chem. Soc. 2023, 145, 1327-1333), in contrast to
experimental observation of rocksalt LiF under ambient conditions. To address
this discrepancy, we employ molecular dynamics (MD) simulations with a refined
machine-learned force field (MLFF), and demonstrate the nucleation and
crystallization of rocksalt LiF from the molten phase at temperatures below
about 800 K. The rocksalt phase remains stable in LiF nanoparticles.
Complementary density functional theory (DFT) calculations show that dispersion
interactions are essential for correctly predicting the thermodynamic stability
of rocksalt LiF over the wurtzite phase on top of the commonly used PBE
functional. Furthermore, we show that inclusion of virial stresses--alongside
energies and forces--in the training of MLFFs is crucial for capturing phase
nucleation and crystallization of rocksalt LiF under the isothermal-isobaric
ensemble. These findings underscore the critical role of dispersion
interactions in atomistic simulations of battery materials, where such effects
are often non-negligible, and highlight the necessity of incorporating virial
stresses during the training of MLFF to enable accurate modeling of solid-state
systems.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 05:52:51 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 08:06:36 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Xu",
"Boyuan",
""
],
[
"Bai",
"Liyi",
""
],
[
"Xu",
"Shenzhen",
""
],
[
"Wu",
"Qisheng",
""
]
] |
2504.12685 | Seungheon Shin | Seungheon Shin, Hridibrata Pal, Jon Pratt, John Niroula, Yinxuan Zhu,
Chandan Joishi, Brianna A. Klein, Andrew Armstrong, Andrew A. Allerman,
Tom\'as Palacios, and Siddharth Rajan | High Breakdown Electric Field (> 5 MV/cm) in UWBG AlGaN Transistors | 14 pages, 10 figures | null | null | null | cond-mat.mtrl-sci physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We report on the design and demonstration of ultra-wide bandgap (UWBG)
AlGaN-channel metal-insulator heterostructure field effect transistors (HEFTs)
for high-power, high-frequency applications. We find that the integration of
gate dielectrics and field plates greatly improves the breakdown field in these
devices, with state-of-art average breakdown field of 5.3 MV/cm (breakdown
voltage > 260 V) with an associated maximum current density of 342 mA/mm, and
cut-off frequency of 9.1 GHz. Furthermore, low trap-related impact was observed
from minimal gate and drain lag estimated from pulsed I-V characteristics. The
reported results provide the potential of UWBG AlGaN HEFTs for the next
generation high-power radio frequency applications.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 06:23:06 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 01:25:56 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Shin",
"Seungheon",
""
],
[
"Pal",
"Hridibrata",
""
],
[
"Pratt",
"Jon",
""
],
[
"Niroula",
"John",
""
],
[
"Zhu",
"Yinxuan",
""
],
[
"Joishi",
"Chandan",
""
],
[
"Klein",
"Brianna A.",
""
],
[
"Armstrong",
"Andrew",
""
],
[
"Allerman",
"Andrew A.",
""
],
[
"Palacios",
"Tomás",
""
],
[
"Rajan",
"Siddharth",
""
]
] |
2504.12735 | Xizhong Guo | Lidong Zhai, Zhijie Qiu, Lvyang Zhang, Jiaqi Li, Yi Wang, Wen Lu,
Xizhong Guo, Ge Sun | The Athenian Academy: A Seven-Layer Architecture Model for Multi-Agent
Systems | null | null | null | null | cs.MA cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper proposes the "Academy of Athens" multi-agent seven-layer
framework, aimed at systematically addressing challenges in multi-agent systems
(MAS) within artificial intelligence (AI) art creation, such as collaboration
efficiency, role allocation, environmental adaptation, and task parallelism.
The framework divides MAS into seven layers: multi-agent collaboration,
single-agent multi-role playing, single-agent multi-scene traversal,
single-agent multi-capability incarnation, different single agents using the
same large model to achieve the same target agent, single-agent using different
large models to achieve the same target agent, and multi-agent synthesis of the
same target agent. Through experimental validation in art creation, the
framework demonstrates its unique advantages in task collaboration, cross-scene
adaptation, and model fusion. This paper further discusses current challenges
such as collaboration mechanism optimization, model stability, and system
security, proposing future exploration through technologies like meta-learning
and federated learning. The framework provides a structured methodology for
multi-agent collaboration in AI art creation and promotes innovative
applications in the art field.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 08:21:28 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 02:45:06 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zhai",
"Lidong",
""
],
[
"Qiu",
"Zhijie",
""
],
[
"Zhang",
"Lvyang",
""
],
[
"Li",
"Jiaqi",
""
],
[
"Wang",
"Yi",
""
],
[
"Lu",
"Wen",
""
],
[
"Guo",
"Xizhong",
""
],
[
"Sun",
"Ge",
""
]
] |
2504.12785 | Davide Liessi | Davide Liessi, Enrico Santi, Rossana Vermiglio, Mayank Thakur, Hil G.
E. Meijer, Francesca Scarabel | New developments in MatCont: delay equation importer and Lyapunov
exponents | submitted to ACM Transactions on Mathematical Software | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | MatCont is a powerful toolbox for numerical bifurcation analysis focussing on
smooth ODEs. A user can study equilibria, periodic and connecting orbits, and
their stability and bifurcations. Here, we report on additional features in
version 7p6. The first is a delay equation importer enabling MatCont users to
study a much larger class of models, namely delay equations with finite delay
(including delay differential and renewal equations). This importer translates
the delay equation into a system of ODEs using a pseudospectral approximation
with an order specified by the user. We also implemented Lyapunov exponent
computations, event functions for Poincar\'e maps, and enhanced homoclinic
continuation. We demonstrate these features with test cases, such as the
Mackey-Glass equation and a renewal equation, and provide additional examples
in online tutorials.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 09:33:27 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 17:40:10 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Liessi",
"Davide",
""
],
[
"Santi",
"Enrico",
""
],
[
"Vermiglio",
"Rossana",
""
],
[
"Thakur",
"Mayank",
""
],
[
"Meijer",
"Hil G. E.",
""
],
[
"Scarabel",
"Francesca",
""
]
] |
2504.12823 | Surbhi Rajput | Surbhi Rajput, Ashish Chiplunkar, Rohit Vaish | Trading Prophets: How to Trade Multiple Stocks Optimally | Published in the SIAM Symposium on Simplicity in Algorithms (SOSA25) | null | null | null | cs.DS cs.GT | http://creativecommons.org/licenses/by/4.0/ | In the single stock trading prophet problem formulated by Correa et al.\
(2023), an online algorithm observes a sequence of prices of a stock. At each
step, the algorithm can either buy the stock by paying the current price if it
doesn't already hold the stock, or it can sell the currently held stock and
collect the current price as a reward. The goal of the algorithm is to maximize
its overall profit.
In this work, we generalize the model and the results of Correa et al.\ by
allowing the algorithm to trade multiple stocks. First, we formulate the
$(k,\ell,\ell')$-Trading Prophet Problem, wherein there are $k$ stocks in the
market, and the online algorithm can hold up to $\ell$ stocks at any time,
where $\ell\leq k$. The online algorithm competes against an offline algorithm
that can hold at most $\ell'\leq\ell$ stocks at any time. Under the assumption
that prices of different stocks are independent, we show that, for any $\ell$,
$\ell'$, and $k$, the optimal competitive ratio of $(k,\ell,\ell')$-Trading
Prophet Problem is $\min(1/2,\ell/k)$.
We further introduce the more general $\cal{M}$-Trading Prophet Problem over
a matroid $\cal{M}$ on the set of $k$ stocks, wherein the stock prices at any
given time are possibly correlated (but are independent across time). The
algorithm is allowed to hold only a feasible subset of stocks at any time. We
prove a tight bound of $1/(1+d)$ on the competitive ratio of the
$\cal{M}$-Trading Prophet Problem, where $d$ is the density of the matroid.
We then consider the non-i.i.d.\ random order setting over a matroid, wherein
stock prices drawn independently from $n$ potentially different distributions
are presented in a uniformly random order. In this setting, we achieve a
competitive ratio of at least $1/(1+d)-\cal{O}(1/n)$, where $d$ is the density
of the matroid, matching the hardness result for i.i.d.\ instances as $n$
approaches $\infty$.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:34:37 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Rajput",
"Surbhi",
""
],
[
"Chiplunkar",
"Ashish",
""
],
[
"Vaish",
"Rohit",
""
]
] |
2504.12828 | Harikrishnan Nellippallil Balakrishnan | Vishrut Ramraj and Nithin Nagaraj and Harikrishnan N B | Predicting Stock Prices using Permutation Decision Trees and Strategic
Trailing | 17 pages, 7 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we explore the application of Permutation Decision Trees (PDT)
and strategic trailing for predicting stock market movements and executing
profitable trades in the Indian stock market. We focus on high-frequency data
using 5-minute candlesticks for the top 50 stocks listed in the NIFTY 50 index.
We implement a trading strategy that aims to buy stocks at lower prices and
sell them at higher prices, capitalizing on short-term market fluctuations. Due
to regulatory constraints in India, short selling is not considered in our
strategy. The model incorporates various technical indicators and employs
hyperparameters such as the trailing stop-loss value and support thresholds to
manage risk effectively. Our results indicate that the proposed trading bot has
the potential to outperform the market average and yield returns higher than
the risk-free rate offered by 10-year Indian government bonds. We trained and
tested data on a 60 day dataset provided by Yahoo Finance. Specifically, 12
days for testing and 48 days for training. Our bot based on permutation
decision tree achieved a profit of 1.3468 % over a 12-day testing period, where
as a bot based on LSTM gave a return of 0.1238 % over a 12-day testing period
and a bot based on RNN gave a return of 0.3096 % over a 12-day testing period.
All of the bots outperform the buy-and-hold strategy, which resulted in a loss
of 2.2508 %.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:42:38 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 07:00:06 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ramraj",
"Vishrut",
""
],
[
"Nagaraj",
"Nithin",
""
],
[
"B",
"Harikrishnan N",
""
]
] |
2504.12843 | Yufan Ge | Francesca Arici, Yufan Ge | Quadratic subproduct systems, free products, and their C*-algebras | null | null | null | null | math.OA math.KT math.QA math.RA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Motivated by the interplay between quadratic algebras, noncommutative
geometry, and operator theory, we introduce the notion of quadratic subproduct
systems of Hilbert spaces. Specifically, we study the subproduct systems
induced by a finite number of complex quadratic polynomials in noncommuting
variables, and describe their Toeplitz and Cuntz--Pimsner algebras. Inspired by
the theory of graded associative algebras, we define a free product operation
in the category of subproduct systems and show that this corresponds to the
reduced free product of the Toeplitz algebras. Finally, we obtain results about
the K-theory of the Toeplitz and Cuntz--Pimsner algebras of a large class of
quadratic subproduct systems.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:57:58 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Arici",
"Francesca",
""
],
[
"Ge",
"Yufan",
""
]
] |
2504.12867 | Guanrou Yang | Guanrou Yang, Chen Yang, Qian Chen, Ziyang Ma, Wenxi Chen, Wen Wang,
Tianrui Wang, Yifan Yang, Zhikang Niu, Wenrui Liu, Fan Yu, Zhihao Du, Zhifu
Gao, ShiLiang Zhang, Xie Chen | EmoVoice: LLM-based Emotional Text-To-Speech Model with Freestyle Text
Prompting | null | null | null | null | eess.AS cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Human speech goes beyond the mere transfer of information; it is a profound
exchange of emotions and a connection between individuals. While Text-to-Speech
(TTS) models have made huge progress, they still face challenges in controlling
the emotional expression in the generated speech. In this work, we propose
EmoVoice, a novel emotion-controllable TTS model that exploits large language
models (LLMs) to enable fine-grained freestyle natural language emotion
control, and a phoneme boost variant design that makes the model output phoneme
tokens and audio tokens in parallel to enhance content consistency, inspired by
chain-of-thought (CoT) and chain-of-modality (CoM) techniques. Besides, we
introduce EmoVoice-DB, a high-quality 40-hour English emotion dataset featuring
expressive speech and fine-grained emotion labels with natural language
descriptions. EmoVoice achieves state-of-the-art performance on the English
EmoVoice-DB test set using only synthetic training data, and on the Chinese
Secap test set using our in-house data. We further investigate the reliability
of existing emotion evaluation metrics and their alignment with human
perceptual preferences, and explore using SOTA multimodal LLMs GPT-4o-audio and
Gemini to assess emotional speech. Demo samples are available at
https://anonymous.4open.science/r/EmoVoice-DF55. Dataset, code, and checkpoints
will be released.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 11:50:04 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 08:18:11 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Guanrou",
""
],
[
"Yang",
"Chen",
""
],
[
"Chen",
"Qian",
""
],
[
"Ma",
"Ziyang",
""
],
[
"Chen",
"Wenxi",
""
],
[
"Wang",
"Wen",
""
],
[
"Wang",
"Tianrui",
""
],
[
"Yang",
"Yifan",
""
],
[
"Niu",
"Zhikang",
""
],
[
"Liu",
"Wenrui",
""
],
[
"Yu",
"Fan",
""
],
[
"Du",
"Zhihao",
""
],
[
"Gao",
"Zhifu",
""
],
[
"Zhang",
"ShiLiang",
""
],
[
"Chen",
"Xie",
""
]
] |
2504.12894 | Mike Roth | Mike Roth | Homeomorphism type of the non-negative part of a complete toric variety | Comments welcome | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this note we show that the nonnegative part of a proper complex toric
variety has the homeomorphism type of a sphere, and consequently that the
nonnegative part has a natural structure of a cell complex. This extends
previous results of Ehlers and Jurkiewicz. The proof also provides a simplicial
decomposition of the nonnegative part, and a parameterization of each maximal
simplex. This result is needed in arXiv:2504.12903 as part of an argument
constructing a torus-stable reduced \v{C}ech complex for any semi-proper toric
variety.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 12:36:31 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 03:34:40 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Roth",
"Mike",
""
]
] |
2504.12903 | Mike Roth | Mike Roth and Sasha Zotine | Reduced \v{C}ech complexes and computing higher direct images under
toric maps | Comments welcome | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper has three main goals :
(1) To give an axiomatic formulation of the construction of "reduced \v{C}ech
complexes", complexes using fewer than the usual number of intersections but
still computing cohomology of an appropriate class of sheaves;
(2) To give a construction of such a reduced \v{C}ech complex for every
semi-proper toric variety $X$, where every open used in the complex is torus
stable, and such that the cell complex governing the reduced \v{C}ech complex
has dimension the cohomological dimension of $X$; and
(3) to give an algorithm to compute the higher direct images of line bundles
relative to a toric fibration between smooth proper toric varieties.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 12:48:18 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 03:39:07 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Roth",
"Mike",
""
],
[
"Zotine",
"Sasha",
""
]
] |
2504.12907 | Andrea Bocchieri Dr | Andrea Bocchieri, Lorenzo V. Mugnai, Enzo Pascale, Andreas
Papageorgiou, Angele Syty, Angelos Tsiaras, Paul Eccleston, Giorgio Savini,
Giovanna Tinetti, Renaud Broquet, Patrick Chapman, and Gianfranco Sechi | De-jittering Ariel: an optimized algorithm | 26 pages, 8 figures, accepted in Experimental Astronomy | null | null | null | astro-ph.EP astro-ph.IM | http://creativecommons.org/licenses/by/4.0/ | The European Space Agency's Ariel mission, scheduled for launch in 2029, aims
to conduct the first large-scale survey of atmospheric spectra of transiting
exoplanets. Ariel achieves the high photometric stability on transit timescales
required to detect the spectroscopic signatures of chemical elements with a
payload design optimized for transit photometry that either eliminates known
systematics or allows for their removal during data processing without
significantly degrading or biasing the detection. Jitter in the spacecraft's
line of sight is a source of disturbance when measuring the spectra of
exoplanet atmospheres. We describe an improved algorithm for de-jittering Ariel
observations simulated in the time domain. We opt for an approach based on the
spatial information on the Point Spread Function (PSF) distortion from jitter
to detrend the optical signals. The jitter model is based on representative
simulations from Airbus Defence and Space, the prime contractor for the Ariel
service module. We investigate the precision and biases of the retrieved
atmospheric spectra from the jitter-detrended observations. At long
wavelengths, the photometric stability of the Ariel spectrometer is already
dominated by photon noise. Our algorithm effectively de-jitters both
photometric and spectroscopic data, ensuring that the performance remains
photon noise-limited across the entire Ariel spectrum, fully compliant with
mission requirements. This work contributes to the development of the data
reduction pipeline for Ariel, aligning with its scientific goals, and may also
benefit other astronomical telescopes and instrumentation.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 12:55:46 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 13:45:10 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Bocchieri",
"Andrea",
""
],
[
"Mugnai",
"Lorenzo V.",
""
],
[
"Pascale",
"Enzo",
""
],
[
"Papageorgiou",
"Andreas",
""
],
[
"Syty",
"Angele",
""
],
[
"Tsiaras",
"Angelos",
""
],
[
"Eccleston",
"Paul",
""
],
[
"Savini",
"Giorgio",
""
],
[
"Tinetti",
"Giovanna",
""
],
[
"Broquet",
"Renaud",
""
],
[
"Chapman",
"Patrick",
""
],
[
"Sechi",
"Gianfranco",
""
]
] |
2504.12943 | Xi Zheng | Xi Zheng, Zhuoyang Li, Xinning Gui, Yuhan Luo | Customizing Emotional Support: How Do Individuals Construct and Interact
With LLM-Powered Chatbots | 20 pages, 3 figures, 3 tables. Accepted to CHI 2025, ACM Conference
on Human Factors in Computing Systems | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Personalized support is essential to fulfill individuals' emotional needs and
sustain their mental well-being. Large language models (LLMs), with great
customization flexibility, hold promises to enable individuals to create their
own emotional support agents. In this work, we developed ChatLab, where users
could construct LLM-powered chatbots with additional interaction features
including voices and avatars. Using a Research through Design approach, we
conducted a week-long field study followed by interviews and design activities
(N = 22), which uncovered how participants created diverse chatbot personas for
emotional reliance, confronting stressors, connecting to intellectual
discourse, reflecting mirrored selves, etc. We found that participants actively
enriched the personas they constructed, shaping the dynamics between themselves
and the chatbot to foster open and honest conversations. They also suggested
other customizable features, such as integrating online activities and
adjustable memory settings. Based on these findings, we discuss opportunities
for enhancing personalized emotional support through emerging AI technologies.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:43:13 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 04:23:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zheng",
"Xi",
""
],
[
"Li",
"Zhuoyang",
""
],
[
"Gui",
"Xinning",
""
],
[
"Luo",
"Yuhan",
""
]
] |
2504.12959 | Dubing Chen | Dubing Chen, Huan Zheng, Jin Fang, Xingping Dong, Xianfei Li, Wenlong
Liao, Tao He, Pai Peng, Jianbing Shen | Rethinking Temporal Fusion with a Unified Gradient Descent View for 3D
Semantic Occupancy Prediction | CVPR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present GDFusion, a temporal fusion method for vision-based 3D semantic
occupancy prediction (VisionOcc). GDFusion opens up the underexplored aspects
of temporal fusion within the VisionOcc framework, focusing on both temporal
cues and fusion strategies. It systematically examines the entire VisionOcc
pipeline, identifying three fundamental yet previously overlooked temporal
cues: scene-level consistency, motion calibration, and geometric
complementation. These cues capture diverse facets of temporal evolution and
make distinct contributions across various modules in the VisionOcc framework.
To effectively fuse temporal signals across heterogeneous representations, we
propose a novel fusion strategy by reinterpreting the formulation of vanilla
RNNs. This reinterpretation leverages gradient descent on features to unify the
integration of diverse temporal information, seamlessly embedding the proposed
temporal cues into the network. Extensive experiments on nuScenes demonstrate
that GDFusion significantly outperforms established baselines. Notably, on
Occ3D benchmark, it achieves 1.4\%-4.8\% mIoU improvements and reduces memory
consumption by 27\%-72\%.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:05:33 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 15:58:20 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Chen",
"Dubing",
""
],
[
"Zheng",
"Huan",
""
],
[
"Fang",
"Jin",
""
],
[
"Dong",
"Xingping",
""
],
[
"Li",
"Xianfei",
""
],
[
"Liao",
"Wenlong",
""
],
[
"He",
"Tao",
""
],
[
"Peng",
"Pai",
""
],
[
"Shen",
"Jianbing",
""
]
] |
2504.12969 | Michael Geynisman | Frederick Schwartz (1), Roberto Acciarri (1), Johan Bremer (2), Roza
Doubnik (1), Caroline Fabre (2), Michael Geynisman (1), Claudio Montanari (1
and 3), Monica Nunes (1), Trevor Nichols (1), William Scofield (1), Zach West
(1), Peter Wilson (1) ((1) Fermi National Accelerator Laboratory, (2) CERN,
(3) INFN, Pavia) | Lessons from commissioning of the cryogenic system for the
Short-Baseline Neutrino Detector at Fermilab | Cryogenics 2023. Proceedings of the 17th IIR International
Conference, Dresden, Germany, april 25-28 2023 | null | 10.18462/iir.cryo.2025.0008 | FERMILAB-CONF-25-0005-PPD | physics.ins-det physics.acc-ph | http://creativecommons.org/licenses/by/4.0/ | Results from commissioning and first year of operations of the cryogenic
system of the Short-Baseline Neutrino Detector (SBND) and its membrane cryostat
installed at the Fermi National Accelerator Laboratory are described. The SBND
detector is installed in a 200 m$^3$ membrane cryostat filled with liquid
argon, which serves both as target and as active media. For the correct
operation of the detector, the liquid argon must be kept in very stable thermal
conditions while the contamination of electronegative impurities must be
consistently kept at the level of small fractions of parts per billion. The
detector is operated in Booster Neutrino Beams (BNB) at Fermilab for the search
of sterile neutrinos and measurements of neutrino-argon cross sections. The
cryostat and the cryogenic systems also serve as prototypes for the much larger
equipment to be used for the LBNF/DUNE experiment. Since its installation in
2018-2023 and cooldown in spring of 2024, the cryostat and the cryogenic system
have been commissioned to support the detector operations. The lessons learned
through installation, testing, commissioning, cooldown, and initial operations
are described.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:19:34 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 14:08:22 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Schwartz",
"Frederick",
"",
"1\n and 3"
],
[
"Acciarri",
"Roberto",
"",
"1\n and 3"
],
[
"Bremer",
"Johan",
"",
"1\n and 3"
],
[
"Doubnik",
"Roza",
"",
"1\n and 3"
],
[
"Fabre",
"Caroline",
"",
"1\n and 3"
],
[
"Geynisman",
"Michael",
"",
"1\n and 3"
],
[
"Montanari",
"Claudio",
"",
"1\n and 3"
],
[
"Nunes",
"Monica",
""
],
[
"Nichols",
"Trevor",
""
],
[
"Scofield",
"William",
""
],
[
"West",
"Zach",
""
],
[
"Wilson",
"Peter",
""
]
] |
2504.12971 | Shiwen Qin | Shiwen Qin, Gabriela Kadlecov\'a, Martin Pil\'at, Shay B. Cohen, Roman
Neruda, Elliot J. Crowley, Jovita Lukasik and Linus Ericsson | Transferrable Surrogates in Expressive Neural Architecture Search Spaces | Project page at: https://shiwenqin.github.io/TransferrableSurrogate/ | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Neural architecture search (NAS) faces a challenge in balancing the
exploration of expressive, broad search spaces that enable architectural
innovation with the need for efficient evaluation of architectures to
effectively search such spaces. We investigate surrogate model training for
improving search in highly expressive NAS search spaces based on context-free
grammars. We show that i) surrogate models trained either using zero-cost-proxy
metrics and neural graph features (GRAF) or by fine-tuning an off-the-shelf LM
have high predictive power for the performance of architectures both within and
across datasets, ii) these surrogates can be used to filter out bad
architectures when searching on novel datasets, thereby significantly speeding
up search and achieving better final performances, and iii) the surrogates can
be further used directly as the search objective for huge speed-ups.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:22:28 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 17:49:14 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Qin",
"Shiwen",
""
],
[
"Kadlecová",
"Gabriela",
""
],
[
"Pilát",
"Martin",
""
],
[
"Cohen",
"Shay B.",
""
],
[
"Neruda",
"Roman",
""
],
[
"Crowley",
"Elliot J.",
""
],
[
"Lukasik",
"Jovita",
""
],
[
"Ericsson",
"Linus",
""
]
] |
2504.12973 | Matthew Citron | Matthew Citron (1), Frank Golf (2), Kranti Gunthoti (3), Andrew Haas
(4), Christopher S. Hill (5), Dariush Imani (6), Samantha Kelly (1), Ming Liu
(3), Steven Lowette (7), Albert De Roeck (8), Sai Neha Santpur (6), Ryan
Schmitz (6), Jacob Steenis (1), David Stuart (6), Yu-Dai Tsai (3), Juan
Salvador Tafoya Vargas (1), Tiepolo Wybouw (7), Jaehyeok Yoo (9) ((1)
University of California, Davis, USA (2) Boston University, Boston, USA (3)
Los Alamos National Laboratory, USA (4) New York University, New York, USA,
(5) The Ohio State University, Columbus, USA, (6) University of California,
Santa Barbara, USA, (7) Vrije Universiteit Brussel, Brussel, Belgium, (8)
CERN, Geneva, Switzerland, (9) Korea University, Seoul, South Korea) | Input to the ESPPU 2026 update: Searching for millicharged particles
with the FORMOSA experiment at the CERN LHC | Contribution prepared for the 2026 update of the European Strategy
for Particle Physics, 9 pages, 6 figures | null | null | null | hep-ex hep-ph physics.ins-det | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this contribution, we evaluate the sensitivity for particles with charges
much smaller than the electron charge with a dedicated scintillator-based
detector in the far forward region at the CERN LHC, FORMOSA. This contribution
will outline the scientific case for this detector, its design and potential
locations, and the sensitivity that can be achieved. The ongoing efforts to
prove the feasibility of the detector with the FORMOSA demonstrator will be
discussed. Finally, possible upgrades to the detector through the use of
high-performance scintillator will be discussed.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:26:33 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Citron",
"Matthew",
""
],
[
"Golf",
"Frank",
""
],
[
"Gunthoti",
"Kranti",
""
],
[
"Haas",
"Andrew",
""
],
[
"Hill",
"Christopher S.",
""
],
[
"Imani",
"Dariush",
""
],
[
"Kelly",
"Samantha",
""
],
[
"Liu",
"Ming",
""
],
[
"Lowette",
"Steven",
""
],
[
"De Roeck",
"Albert",
""
],
[
"Santpur",
"Sai Neha",
""
],
[
"Schmitz",
"Ryan",
""
],
[
"Steenis",
"Jacob",
""
],
[
"Stuart",
"David",
""
],
[
"Tsai",
"Yu-Dai",
""
],
[
"Vargas",
"Juan Salvador Tafoya",
""
],
[
"Wybouw",
"Tiepolo",
""
],
[
"Yoo",
"Jaehyeok",
""
]
] |
2504.12990 | Matteo Becchi | Matteo Becchi (1), Giovanni Maria Pavan (1) ((1) Politecnico di
Torino, Dipartimento di Scienze Applicate e Tecnologia) | Maximum Information Extraction From Noisy Data Via Shannon Entropy
Minimization | Main text 7 pages, 3 figures; Supplemental Materials 3 pages, 3
figures. v2: author's email was missing | null | null | null | physics.data-an | http://creativecommons.org/licenses/by/4.0/ | Granting maximum information extraction in the analysis of noisy data is
non-trivial. We introduce a general, data-driven approach that employs Shannon
entropy as a transferable metric to quantify the maximum information
extractable from noisy data via their clustering into statistically-relevant
micro-domains. We demonstrate the method's efficiency by analyzing, as a
representative example, time-series data extracted from molecular dynamics
simulations of water and ice coexisting at the solid/liquid transition
temperature. The method allows quantifying the information contained in the
data distributions (time-independent component) and the additional information
gain attainable by analyzing data as time-series (i.e., accounting for the
information contained in data time-correlations). The approach is also highly
effective for high-dimensional datasets, providing clear demonstrations of how
considering components/data that may be little informative but noisy may be not
only useless but even detrimental to maximum information extraction. This
provides a general and robust parameter-free approach and quantitative metrics
for data-analysis, and for the study of any type of system from its data.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:54:46 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 12:44:56 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Becchi",
"Matteo",
""
],
[
"Pavan",
"Giovanni Maria",
""
]
] |
2504.13014 | Tania Robens | Tania Robens | Extended scalar sectors from all angles (in 15 minutes) | 6 pages, 4 figures; contribution to the Proceedings of 59th
Rencontres de Moriond 2025, Electroweak Interactions & Unified Theories; v2:
minor changes | null | null | RBI-ThPhys-2025-21, CERN-TH-2025-083 | hep-ph | http://creativecommons.org/licenses/by/4.0/ | In this proceedings contribution, I briefly summarize various aspects that
are important in the discussions of new physics searches with novel scalar
states, at current and future colliders. In particular, I give a brief glance
on the status of two Higgs doublet models, and discuss multi-scalar production
as well as interference effects in Di-Higgs searches. I also mention searches
of new scalar final states at possible Higgs factories.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 15:19:24 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 08:33:11 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Robens",
"Tania",
""
]
] |
2504.13037 | Yundi Zhang | Yundi Zhang, Paul Hager, Che Liu, Suprosanna Shit, Chen Chen, Daniel
Rueckert, Jiazhen Pan | Towards Cardiac MRI Foundation Models: Comprehensive Visual-Tabular
Representations for Whole-Heart Assessment and Beyond | null | null | null | null | eess.IV cs.AI cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cardiac magnetic resonance imaging is the gold standard for non-invasive
cardiac assessment, offering rich spatio-temporal views of the cardiac anatomy
and physiology. Patient-level health factors, such as demographics, metabolic,
and lifestyle, are known to substantially influence cardiovascular health and
disease risk, yet remain uncaptured by CMR alone. To holistically understand
cardiac health and to enable the best possible interpretation of an
individual's disease risk, CMR and patient-level factors must be jointly
exploited within an integrated framework. Recent multi-modal approaches have
begun to bridge this gap, yet they often rely on limited spatio-temporal data
and focus on isolated clinical tasks, thereby hindering the development of a
comprehensive representation for cardiac health evaluation. To overcome these
limitations, we introduce ViTa, a step toward foundation models that delivers a
comprehensive representation of the heart and a precise interpretation of
individual disease risk. Leveraging data from 42,000 UK Biobank participants,
ViTa integrates 3D+T cine stacks from short-axis and long-axis views, enabling
a complete capture of the cardiac cycle. These imaging data are then fused with
detailed tabular patient-level factors, enabling context-aware insights. This
multi-modal paradigm supports a wide spectrum of downstream tasks, including
cardiac phenotype and physiological feature prediction, segmentation, and
classification of cardiac and metabolic diseases within a single unified
framework. By learning a shared latent representation that bridges rich imaging
features and patient context, ViTa moves beyond traditional, task-specific
models toward a universal, patient-specific understanding of cardiac health,
highlighting its potential to advance clinical utility and scalability in
cardiac analysis.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 15:46:19 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 09:26:55 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zhang",
"Yundi",
""
],
[
"Hager",
"Paul",
""
],
[
"Liu",
"Che",
""
],
[
"Shit",
"Suprosanna",
""
],
[
"Chen",
"Chen",
""
],
[
"Rueckert",
"Daniel",
""
],
[
"Pan",
"Jiazhen",
""
]
] |
2504.13042 | Dachun Kai | Dachun Kai, Yueyi Zhang, Jin Wang, Zeyu Xiao, Zhiwei Xiong, Xiaoyan
Sun | Event-Enhanced Blurry Video Super-Resolution | AAAI 2025. Project page:
https://dachunkai.github.io/ev-deblurvsr.github.io/ | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we tackle the task of blurry video super-resolution (BVSR),
aiming to generate high-resolution (HR) videos from low-resolution (LR) and
blurry inputs. Current BVSR methods often fail to restore sharp details at high
resolutions, resulting in noticeable artifacts and jitter due to insufficient
motion information for deconvolution and the lack of high-frequency details in
LR frames. To address these challenges, we introduce event signals into BVSR
and propose a novel event-enhanced network, Ev-DeblurVSR. To effectively fuse
information from frames and events for feature deblurring, we introduce a
reciprocal feature deblurring module that leverages motion information from
intra-frame events to deblur frame features while reciprocally using global
scene context from the frames to enhance event features. Furthermore, to
enhance temporal consistency, we propose a hybrid deformable alignment module
that fully exploits the complementary motion information from inter-frame
events and optical flow to improve motion estimation in the deformable
alignment process. Extensive evaluations demonstrate that Ev-DeblurVSR
establishes a new state-of-the-art performance on both synthetic and real-world
datasets. Notably, on real data, our method is +2.59 dB more accurate and
7.28$\times$ faster than the recent best BVSR baseline FMA-Net. Code:
https://github.com/DachunKai/Ev-DeblurVSR.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 15:55:41 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 02:49:30 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kai",
"Dachun",
""
],
[
"Zhang",
"Yueyi",
""
],
[
"Wang",
"Jin",
""
],
[
"Xiao",
"Zeyu",
""
],
[
"Xiong",
"Zhiwei",
""
],
[
"Sun",
"Xiaoyan",
""
]
] |
2504.13074 | Guibin Chen | Guibin Chen, Dixuan Lin, Jiangping Yang, Chunze Lin, Juncheng Zhu,
Mingyuan Fan, Hao Zhang, Sheng Chen, Zheng Chen, Chengchen Ma, Weiming Xiong,
Wei Wang, Nuo Pang, Kang Kang, Zhiheng Xu, Yuzhe Jin, Yupeng Liang, Yubing
Song, Peng Zhao, Boyuan Xu, Di Qiu, Debang Li, Zhengcong Fei, Yang Li, Yahui
Zhou | SkyReels-V2: Infinite-length Film Generative Model | 31 pages,10 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Recent advances in video generation have been driven by diffusion models and
autoregressive frameworks, yet critical challenges persist in harmonizing
prompt adherence, visual quality, motion dynamics, and duration: compromises in
motion dynamics to enhance temporal visual quality, constrained video duration
(5-10 seconds) to prioritize resolution, and inadequate shot-aware generation
stemming from general-purpose MLLMs' inability to interpret cinematic grammar,
such as shot composition, actor expressions, and camera motions. These
intertwined limitations hinder realistic long-form synthesis and professional
film-style generation. To address these limitations, we propose SkyReels-V2, an
Infinite-length Film Generative Model, that synergizes Multi-modal Large
Language Model (MLLM), Multi-stage Pretraining, Reinforcement Learning, and
Diffusion Forcing Framework. Firstly, we design a comprehensive structural
representation of video that combines the general descriptions by the
Multi-modal LLM and the detailed shot language by sub-expert models. Aided with
human annotation, we then train a unified Video Captioner, named
SkyCaptioner-V1, to efficiently label the video data. Secondly, we establish
progressive-resolution pretraining for the fundamental video generation,
followed by a four-stage post-training enhancement: Initial concept-balanced
Supervised Fine-Tuning (SFT) improves baseline quality; Motion-specific
Reinforcement Learning (RL) training with human-annotated and synthetic
distortion data addresses dynamic artifacts; Our diffusion forcing framework
with non-decreasing noise schedules enables long-video synthesis in an
efficient search space; Final high-quality SFT refines visual fidelity. All the
code and models are available at https://github.com/SkyworkAI/SkyReels-V2.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:37:27 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 09:46:32 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Chen",
"Guibin",
""
],
[
"Lin",
"Dixuan",
""
],
[
"Yang",
"Jiangping",
""
],
[
"Lin",
"Chunze",
""
],
[
"Zhu",
"Juncheng",
""
],
[
"Fan",
"Mingyuan",
""
],
[
"Zhang",
"Hao",
""
],
[
"Chen",
"Sheng",
""
],
[
"Chen",
"Zheng",
""
],
[
"Ma",
"Chengchen",
""
],
[
"Xiong",
"Weiming",
""
],
[
"Wang",
"Wei",
""
],
[
"Pang",
"Nuo",
""
],
[
"Kang",
"Kang",
""
],
[
"Xu",
"Zhiheng",
""
],
[
"Jin",
"Yuzhe",
""
],
[
"Liang",
"Yupeng",
""
],
[
"Song",
"Yubing",
""
],
[
"Zhao",
"Peng",
""
],
[
"Xu",
"Boyuan",
""
],
[
"Qiu",
"Di",
""
],
[
"Li",
"Debang",
""
],
[
"Fei",
"Zhengcong",
""
],
[
"Li",
"Yang",
""
],
[
"Zhou",
"Yahui",
""
]
] |
2504.13127 | Jake Ketchum | Jake Ketchum, James Avtges, Millicent Schlafly, Helena Young,
Taekyoung Kim, Ryan L. Truby, Todd D. Murphey | Force and Speed in a Soft Stewart Platform | Published at Robosoft 2025 | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many soft robots struggle to produce dynamic motions with fast, large
displacements. We develop a parallel 6 degree-of-freedom (DoF) Stewart-Gough
mechanism using Handed Shearing Auxetic (HSA) actuators. By using soft
actuators, we are able to use one third as many mechatronic components as a
rigid Stewart platform, while retaining a working payload of 2kg and an
open-loop bandwidth greater than 16Hz. We show that the platform is capable of
both precise tracing and dynamic disturbance rejection when controlling a ball
and sliding puck using a Proportional Integral Derivative (PID) controller. We
develop a machine-learning-based kinematics model and demonstrate a functional
workspace of roughly 10cm in each translation direction and 28 degrees in each
orientation. This 6DoF device has many of the characteristics associated with
rigid components - power, speed, and total workspace - while capturing the
advantages of soft mechanisms.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:43:33 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 15:12:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ketchum",
"Jake",
""
],
[
"Avtges",
"James",
""
],
[
"Schlafly",
"Millicent",
""
],
[
"Young",
"Helena",
""
],
[
"Kim",
"Taekyoung",
""
],
[
"Truby",
"Ryan L.",
""
],
[
"Murphey",
"Todd D.",
""
]
] |
2504.13135 | Rui-Qing Xiao | John Ellis, Hong-Jian He, Rui-Qing Xiao | Probing CP-Violating Neutral Triple Gauge Couplings at Electron-Positron
Colliders | null | null | null | KCL-PH-TH/2025-01, CERN-TH-2025-002 | hep-ph hep-ex | http://creativecommons.org/licenses/by/4.0/ | We study the forms of CP-violating (CPV) neutral triple gauge couplings
(nTGCs) that can be realized via dimension-8 operators in the Standard Model
Effective Field Theory (SMEFT). We present a new formulation of the CPV nTGC
form factors that is compatible with the spontaneous breaking of the
electroweak gauge symmetry, and show how these CPV form factors can be matched
consistently with the corresponding dimension-8 CPV nTGC operators in the
broken phase. We then study probes of the CPV nTGCs at future high-energy
$e^+e^-$ colliders with centre-of-mass energies $\sqrt{s} = (0.25, 0.5,1, 3,
5)$TeV respectively, demonstrating that the $e^{\mp}$ beam polarizations can
help to improve the sensitivities of probes of the nTGCs. We estimate that the
sensitivities for probing the new physics scales of the nTGCs can range from
${O}(\rm{TeV})$ at a 250GeV $e^+e^-$ collider to ${O}(10\rm{TeV})$ at a 5TeV
$e^+e^-$ collider, and that the sensitivities to form factors range from
${O}(10^{-4})$ to ${O}(10^{-8})$.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:47:54 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ellis",
"John",
""
],
[
"He",
"Hong-Jian",
""
],
[
"Xiao",
"Rui-Qing",
""
]
] |
2504.13167 | Zetong Zhang | Zetong Zhang, Manuel Kaufmann, Lixin Xue, Jie Song, Martin R. Oswald | ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from
Monocular Videos | Accepted at CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Creating a photorealistic scene and human reconstruction from a single
monocular in-the-wild video figures prominently in the perception of a
human-centric 3D world. Recent neural rendering advances have enabled holistic
human-scene reconstruction but require pre-calibrated camera and human poses,
and days of training time. In this work, we introduce a novel unified framework
that simultaneously performs camera tracking, human pose estimation and
human-scene reconstruction in an online fashion. 3D Gaussian Splatting is
utilized to learn Gaussian primitives for humans and scenes efficiently, and
reconstruction-based camera tracking and human pose estimation modules are
designed to enable holistic understanding and effective disentanglement of pose
and appearance. Specifically, we design a human deformation module to
reconstruct the details and enhance generalizability to out-of-distribution
poses faithfully. Aiming to learn the spatial correlation between human and
scene accurately, we introduce occlusion-aware human silhouette rendering and
monocular geometric priors, which further improve reconstruction quality.
Experiments on the EMDB and NeuMan datasets demonstrate superior or on-par
performance with existing methods in camera tracking, human pose estimation,
novel view synthesis and runtime. Our project page is at
https://eth-ait.github.io/ODHSR.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:59:02 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 17:00:33 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zhang",
"Zetong",
""
],
[
"Kaufmann",
"Manuel",
""
],
[
"Xue",
"Lixin",
""
],
[
"Song",
"Jie",
""
],
[
"Oswald",
"Martin R.",
""
]
] |
2504.13182 | Michael Lopez II | Michael T. Lopez II | Challenging the Eye-Mind Link Hypothesis: Visualizing Gazes For Each
Programming Problem | Accepted as poster paper at the 32nd International Conference on
Computers in Education (ICCE 2024) | null | null | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This investigates the relationship between eye fixation patterns and
performance in Java programming exercises using eye-tracking technology.
Thirty-one students from a university in Metro Manila participated, and their
eye movements were recorded while solving five Java programming exercises
(three of the five exercises were picked). The fixation data were preprocessed
and visualized using heatmap bin graphs, dividing the participants into correct
and wrong answer groups. The Mann-Whitney U Test was employed to determine if
there were significant differences in the fixation patterns between the two
groups.
| [
{
"version": "v1",
"created": "Sun, 12 Jan 2025 17:18:54 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Lopez",
"Michael T.",
"II"
]
] |
2504.13183 | Rawan AlMakinah | Rawan AlMakinah | Factors That Influence the Adoption of AI-enabled Conversational Agents
(AICAs) as an Augmenting Therapeutic Tool by Frontline Healthcare Workers:
From Technology Acceptance Model 3 (TAM3) Lens -- A Systematic Mapping Review | null | null | null | null | cs.HC cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Artificial intelligent (AI) conversational agents hold a promising future in
the field of mental health, especially in helping marginalized communities that
lack access to mental health support services. It is tempting to have a 24/7
mental health companion that can be accessed anywhere using mobile phones to
provide therapist-like advice. Yet, caution should be taken, and studies around
their feasibility need to be surveyed. Before adopting such a rapidly changing
technology, studies on its feasibility should be explored, summarized, and
synthesized to gain a solid understanding of the status quo and to enable us to
build a framework that can guide us throughout the development and deployment
processes. Different perspectives must be considered when investigating the
feasibility of AI conversational agents, including the mental healthcare
professional perspective. The literature can provide insights into their
perspectives in terms of opportunities, concerns, and implications. Mental
health professionals, the subject-matter experts in this field, have their
points of view that should be understood and considered. This systematic
literature review will explore mental health practitioners' attitudes toward AI
conversational agents and the factors that affect their adoption and
recommendation of the technology to augment their services and treatments. The
TAM3 Framework will be the lens through which this systematic literature review
will be conducted.
| [
{
"version": "v1",
"created": "Sun, 26 Jan 2025 02:31:27 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"AlMakinah",
"Rawan",
""
]
] |
2504.13184 | Youning Li | Youning Li | Understanding the Optical Theorem of Scattering: Scattering Surface Area
against Scattering Cross Section, an Example with Ellipsoidal Scattering | 8 pages, 2 figures, comments welcome | EPL, 149 (2025) 30003 | 10.1209/0295-5075/adab15 | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | In this paper, we propose using the scattering surface area rather than the
scattering cross section to characterize the scattering behavior of ellipsoidal
rigid bodies. We examined the scattering behavior of ellipsoidal rigid bodies,
focusing on the relationship between their surface area and total scattering
cross-section. Building on the foundational work of Carson Flammer, we utilize
the spheroidal coordinate system to derive solutions for both prolate and
oblate spheroids. Our analysis reveals that under the long-wavelength
approximation, the total scattering cross-section is equivalent to the surface
area of the ellipsoid, a relationship that holds true for both small and
moderate eccentricities. This finding extends the established optical theorem,
previously validated for spherical bodies, to more complex geometries.
| [
{
"version": "v1",
"created": "Sun, 16 Mar 2025 14:41:42 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Li",
"Youning",
""
]
] |
2504.13185 | Guilherme Xavier | Daniel Spegel-Lexne, Jo\~ao Manoel Barbosa Pereira, Alvaro Alarc\'on,
Joakim Argillander, Martin Clason, {\AA}sa Claesson, Kenny Hey Tow, Walter
Margulis, Guilherme B. Xavier | Storage Buffer of Polarization Quantum States Based on a Poled-Fiber
Phase Modulator | 2 pages, 2 figures. Conference paper, accepted at CLEO 2025 | null | null | null | quant-ph physics.optics | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Dynamic storage of qubits is crucial for quantum communication networks. Here
we present an adjustable buffer capable of storing photonic polarization
quantum states in a fiber loop controllable by a poled fiber modulator.
| [
{
"version": "v1",
"created": "Thu, 20 Mar 2025 19:55:29 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Spegel-Lexne",
"Daniel",
""
],
[
"Pereira",
"João Manoel Barbosa",
""
],
[
"Alarcón",
"Alvaro",
""
],
[
"Argillander",
"Joakim",
""
],
[
"Clason",
"Martin",
""
],
[
"Claesson",
"Åsa",
""
],
[
"Tow",
"Kenny Hey",
""
],
[
"Margulis",
"Walter",
""
],
[
"Xavier",
"Guilherme B.",
""
]
] |
2504.13186 | Hamza Kheddar | Yassine Habchi, Hamza Kheddar, Yassine Himeur, Adel Belouchrani,
Erchin Serpedin, Fouad Khelifi, Muhammad E.H. Chowdhury | Advanced Deep Learning and Large Language Models: Comprehensive Insights
for Cancer Detection | null | Image and Vision Computing, Elsevier, 2025 | 10.1016/j.imavis.2025.105495 | null | eess.IV cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | The rapid advancement of deep learning (DL) has transformed healthcare,
particularly in cancer detection and diagnosis. DL surpasses traditional
machine learning and human accuracy, making it a critical tool for identifying
diseases. Despite numerous reviews on DL in healthcare, a comprehensive
analysis of its role in cancer detection remains limited. Existing studies
focus on specific aspects, leaving gaps in understanding its broader impact.
This paper addresses these gaps by reviewing advanced DL techniques, including
transfer learning (TL), reinforcement learning (RL), federated learning (FL),
Transformers, and large language models (LLMs). These approaches enhance
accuracy, tackle data scarcity, and enable decentralized learning while
maintaining data privacy. TL adapts pre-trained models to new datasets,
improving performance with limited labeled data. RL optimizes diagnostic
pathways and treatment strategies, while FL fosters collaborative model
development without sharing sensitive data. Transformers and LLMs,
traditionally used in natural language processing, are now applied to medical
data for improved interpretability. Additionally, this review examines these
techniques' efficiency in cancer diagnosis, addresses challenges like data
imbalance, and proposes solutions. It serves as a resource for researchers and
practitioners, providing insights into current trends and guiding future
research in advanced DL for cancer detection.
| [
{
"version": "v1",
"created": "Sun, 30 Mar 2025 15:17:40 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Habchi",
"Yassine",
""
],
[
"Kheddar",
"Hamza",
""
],
[
"Himeur",
"Yassine",
""
],
[
"Belouchrani",
"Adel",
""
],
[
"Serpedin",
"Erchin",
""
],
[
"Khelifi",
"Fouad",
""
],
[
"Chowdhury",
"Muhammad E. H.",
""
]
] |
2504.13187 | In Hak Moon | In Hak Moon | Benchmarking Large Language Models for Calculus Problem-Solving: A
Comparative Analysis | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | This study presents a comprehensive evaluation of five leading large language
models (LLMs) - Chat GPT 4o, Copilot Pro, Gemini Advanced, Claude Pro, and Meta
AI - on their performance in solving calculus differentiation problems. The
investigation assessed these models across 13 fundamental problem types,
employing a systematic cross-evaluation framework where each model solved
problems generated by all models. Results revealed significant performance
disparities, with Chat GPT 4o achieving the highest success rate (94.71%),
followed by Claude Pro (85.74%), Gemini Advanced (84.42%), Copilot Pro
(76.30%), and Meta AI (56.75%). All models excelled at procedural
differentiation tasks but showed varying limitations with conceptual
understanding and algebraic manipulation. Notably, problems involving
increasing/decreasing intervals and optimization word problems proved most
challenging across all models. The cross-evaluation matrix revealed that Claude
Pro generated the most difficult problems, suggesting distinct capabilities
between problem generation and problem-solving. These findings have significant
implications for educational applications, highlighting both the potential and
limitations of LLMs as calculus learning tools. While they demonstrate
impressive procedural capabilities, their conceptual understanding remains
limited compared to human mathematical reasoning, emphasizing the continued
importance of human instruction for developing deeper mathematical
comprehension.
| [
{
"version": "v1",
"created": "Mon, 31 Mar 2025 00:39:40 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Moon",
"In Hak",
""
]
] |
2504.13188 | Takayuki Ishitobi | Takayuki Ishitobi | Landau theory of CeCoSi | 2 pages, 1 figure | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | We investigate both the nonmagnetic and magnetic ordered phases of CeCoSi
using Landau theory. Our analysis predicts that three successive phase
transitions occur at zero magnetic field. In the intermediate phase, an
out-of-plane magnetic moment may be present or absent. If present, magnetic
fields applied along the [100] and [110] directions induce additional magnetic
phases.
| [
{
"version": "v1",
"created": "Tue, 1 Apr 2025 08:36:10 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ishitobi",
"Takayuki",
""
]
] |
2504.13189 | Sohom Ghosh | Sohom Ghosh and Sudip Kumar Naskar | BASIR: Budget-Assisted Sectoral Impact Ranking -- A Dataset for Sector
Identification and Performance Prediction Using Language Models | The codes and the datasets can be accessed from
https://huggingface.co/datasets/sohomghosh/BASIR_Budget_Assisted_Sectoral_Impact_Ranking/tree/main/ | null | null | null | cs.CL q-fin.ST | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Government fiscal policies, particularly annual union budgets, exert
significant influence on financial markets. However, real-time analysis of
budgetary impacts on sector-specific equity performance remains
methodologically challenging and largely unexplored. This study proposes a
framework to systematically identify and rank sectors poised to benefit from
India's Union Budget announcements. The framework addresses two core tasks: (1)
multi-label classification of excerpts from budget transcripts into 81
predefined economic sectors, and (2) performance ranking of these sectors.
Leveraging a comprehensive corpus of Indian Union Budget transcripts from 1947
to 2025, we introduce BASIR (Budget-Assisted Sectoral Impact Ranking), an
annotated dataset mapping excerpts from budgetary transcripts to sectoral
impacts. Our architecture incorporates fine-tuned embeddings for sector
identification, coupled with language models that rank sectors based on their
predicted performances. Our results demonstrate 0.605 F1-score in sector
classification, and 0.997 NDCG score in predicting ranks of sectors based on
post-budget performances. The methodology enables investors and policymakers to
quantify fiscal policy impacts through structured, data-driven insights,
addressing critical gaps in manual analysis. The annotated dataset has been
released under CC-BY-NC-SA-4.0 license to advance computational economics
research.
| [
{
"version": "v1",
"created": "Wed, 2 Apr 2025 13:10:26 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ghosh",
"Sohom",
""
],
[
"Naskar",
"Sudip Kumar",
""
]
] |
2504.13190 | Yuyang Du | Liujianfu Wang, Xinyi Long, Yuyang Du, Xiaoyan Liu, Kexin Chen, Soung
Chang Liew | Cellular-X: An LLM-empowered Cellular Agent for Efficient Base Station
Operations | MobiSys â25, June 23-27, 2025, Anaheim, CA, USA | null | null | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper introduces Cellular-X, an LLM-powered agent designed to automate
cellular base station (BS) maintenance. Leveraging multimodal LLM and
retrieval-augmented generation (RAG) techniques, Cellular-X significantly
enhances field engineer efficiency by quickly interpreting user intents,
retrieving relevant technical information, and configuring a BS through
iterative self-correction. Key features of the demo include automatic
customized BS setup, document-based query answering, and voice-controlled
configuration reporting and revision. We implemented Cellular-X on a USRP X310
testbed for demonstration. Demo videos and implementation details are available
at https://github.com/SeaBreezing/Cellular-X.
| [
{
"version": "v1",
"created": "Thu, 10 Apr 2025 19:04:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Wang",
"Liujianfu",
""
],
[
"Long",
"Xinyi",
""
],
[
"Du",
"Yuyang",
""
],
[
"Liu",
"Xiaoyan",
""
],
[
"Chen",
"Kexin",
""
],
[
"Liew",
"Soung Chang",
""
]
] |
2504.13191 | Nam Nguyen | Nam Nguyen | Universal Representations for Classification-enhanced Lossy Compression | null | null | null | null | cs.CV cs.AI cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In lossy compression, the classical tradeoff between compression rate and
reconstruction distortion has traditionally guided algorithm design. However,
Blau and Michaeli [5] introduced a generalized framework, known as the
rate-distortion-perception (RDP) function, incorporating perceptual quality as
an additional dimension of evaluation. More recently, the
rate-distortion-classification (RDC) function was investigated in [19],
evaluating compression performance by considering classification accuracy
alongside distortion. In this paper, we explore universal representations,
where a single encoder is developed to achieve multiple decoding objectives
across various distortion and classification (or perception) constraints. This
universality avoids retraining encoders for each specific operating point
within these tradeoffs. Our experimental validation on the MNIST dataset
indicates that a universal encoder incurs only minimal performance degradation
compared to individually optimized encoders for perceptual image compression
tasks, aligning with prior results from [23]. Nonetheless, we also identify
that in the RDC setting, reusing an encoder optimized for one specific
classification-distortion tradeoff leads to a significant distortion penalty
when applied to alternative points.
| [
{
"version": "v1",
"created": "Sat, 12 Apr 2025 00:55:56 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Nguyen",
"Nam",
""
]
] |
2504.13192 | Liangbo Ning | Liang-bo Ning, Shijie Wang, Wenqi Fan, Qing Li, Xin Xu, Hao Chen,
Feiran Huang | CheatAgent: Attacking LLM-Empowered Recommender Systems via LLM Agent | null | null | 10.1145/3637528.3671837 | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, Large Language Model (LLM)-empowered recommender systems (RecSys)
have brought significant advances in personalized user experience and have
attracted considerable attention. Despite the impressive progress, the research
question regarding the safety vulnerability of LLM-empowered RecSys still
remains largely under-investigated. Given the security and privacy concerns, it
is more practical to focus on attacking the black-box RecSys, where attackers
can only observe the system's inputs and outputs. However, traditional attack
approaches employing reinforcement learning (RL) agents are not effective for
attacking LLM-empowered RecSys due to the limited capabilities in processing
complex textual inputs, planning, and reasoning. On the other hand, LLMs
provide unprecedented opportunities to serve as attack agents to attack RecSys
because of their impressive capability in simulating human-like decision-making
processes. Therefore, in this paper, we propose a novel attack framework called
CheatAgent by harnessing the human-like capabilities of LLMs, where an
LLM-based agent is developed to attack LLM-Empowered RecSys. Specifically, our
method first identifies the insertion position for maximum impact with minimal
input modification. After that, the LLM agent is designed to generate
adversarial perturbations to insert at target positions. To further improve the
quality of generated perturbations, we utilize the prompt tuning technique to
improve attacking strategies via feedback from the victim RecSys iteratively.
Extensive experiments across three real-world datasets demonstrate the
effectiveness of our proposed attacking method.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 05:31:37 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ning",
"Liang-bo",
""
],
[
"Wang",
"Shijie",
""
],
[
"Fan",
"Wenqi",
""
],
[
"Li",
"Qing",
""
],
[
"Xu",
"Xin",
""
],
[
"Chen",
"Hao",
""
],
[
"Huang",
"Feiran",
""
]
] |
2504.13193 | Hong Yang | Hong Yang | HEAT:History-Enhanced Dual-phase Actor-Critic Algorithm with A Shared
Transformer | null | null | null | null | cs.NI cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For a single-gateway LoRaWAN network, this study proposed a history-enhanced
two-phase actor-critic algorithm with a shared transformer algorithm (HEAT) to
improve network performance. HEAT considers uplink parameters and often
neglected downlink parameters, and effectively integrates offline and online
reinforcement learning, using historical data and real-time interaction to
improve model performance. In addition, this study developed an open source
LoRaWAN network simulator LoRaWANSim. The simulator considers the demodulator
lock effect and supports multi-channel, multi-demodulator and bidirectional
communication. Simulation experiments show that compared with the best results
of all compared algorithms, HEAT improves the packet success rate and energy
efficiency by 15% and 95%, respectively.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 10:32:02 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Hong",
""
]
] |
2504.13194 | Hong Yang | Hong Yang | Optimizing Multi-Gateway LoRaWAN via Cloud-Edge Collaboration and
Knowledge Distillation | null | null | null | null | cs.NI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For large-scale multi-gateway LoRaWAN networks, this study proposes a
cloud-edge collaborative resource allocation and decision-making method based
on edge intelligence, HEAT-LDL (HEAT-Local Distill Lyapunov), which realizes
collaborative decision-making between gateways and terminal nodes. HEAT-LDL
combines the Actor-Critic architecture and the Lyapunov optimization method to
achieve intelligent downlink control and gateway load balancing. When the
signal quality is good, the network server uses the HEAT algorithm to schedule
the terminal nodes. To improve the efficiency of autonomous decision-making of
terminal nodes, HEAT-LDL performs cloud-edge knowledge distillation on the HEAT
teacher model on the terminal node side. When the downlink decision instruction
is lost, the terminal node uses the student model and the edge decider based on
prior knowledge and local history to make collaborative autonomous decisions.
Simulation experiments show that compared with the optimal results of all
compared algorithms, HEAT-LDL improves the packet success rate and energy
efficiency by 20.5% and 88.1%, respectively.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 10:34:37 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Hong",
""
]
] |
2504.13195 | Runbo Li | Runbo Li | On prime-producing sieves and distribution of $\alpha p-\beta$ mod $1$ | 17 pages | null | null | null | math.NT | http://creativecommons.org/licenses/by-sa/4.0/ | The author proves that there are infinitely many primes $p$ such that $\|
\alpha p - \beta \| < p^{-\frac{28}{87}}$, where $\alpha$ is an irrational
number and $\beta$ is a real number. This sharpens a result of Jia (2000) and
provides a new triple $(\gamma, \theta, \nu)=(\frac{59}{87}, \frac{28}{87},
\frac{1}{29})$ that can produce primes in Ford and Maynard's work on
prime-producing sieves. Our minimum amount of Type-II information required
($\nu = \frac{1}{29}$) is less than any previous work on this topic using only
traditional Type-I and Type-II information.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 03:40:03 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Li",
"Runbo",
""
]
] |
2504.13196 | Leonid Legashev | Leonid Legashev and Arthur Zhigalov | Investigating cybersecurity incidents using large language models in
latest-generation wireless networks | 11 pages, 2 figures | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | The purpose of research: Detection of cybersecurity incidents and analysis of
decision support and assessment of the effectiveness of measures to counter
information security threats based on modern generative models. The methods of
research: Emulation of signal propagation data in MIMO systems, synthesis of
adversarial examples, execution of adversarial attacks on machine learning
models, fine tuning of large language models for detecting adversarial attacks,
explainability of decisions on detecting cybersecurity incidents based on the
prompts technique. Scientific novelty: A binary classification of data
poisoning attacks was performed using large language models, and the
possibility of using large language models for investigating cybersecurity
incidents in the latest generation wireless networks was investigated. The
result of research: Fine-tuning of large language models was performed on the
prepared data of the emulated wireless network segment. Six large language
models were compared for detecting adversarial attacks, and the capabilities of
explaining decisions made by a large language model were investigated. The
Gemma-7b model showed the best results according to the metrics Precision =
0.89, Recall = 0.89 and F1-Score = 0.89. Based on various explainability
prompts, the Gemma-7b model notes inconsistencies in the compromised data under
study, performs feature importance analysis and provides various
recommendations for mitigating the consequences of adversarial attacks. Large
language models integrated with binary classifiers of network threats have
significant potential for practical application in the field of cybersecurity
incident investigation, decision support and assessing the effectiveness of
measures to counter information security threats.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 09:57:20 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Legashev",
"Leonid",
""
],
[
"Zhigalov",
"Arthur",
""
]
] |
2504.13197 | Takayuki Ishitobi | Takayuki Ishitobi | Superzone gap formation induced by ferroic orders | 2 pages, 2 figures | null | null | null | cond-mat.str-el cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | We demonstrate that a superzone gap, typically associated with antiferroic
ordering, can also emerge from ferroic orders in systems with sublattice
degrees of freedom. By analyzing a $p$-orbital tight-binding model on a zigzag
chain, we show that a Su--Schrieffer--Heeger-type gap is induced by
ferroquadrupolar or ferromagnetic order or by applying an external magnetic
field.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 11:22:50 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ishitobi",
"Takayuki",
""
]
] |
2504.13198 | Eric Landquist | Eric Landquist and Nimit Sawhney and Simer Sawhney | Overcoming Bottlenecks in Homomorphic Encryption for the 2024 Mexican
Federal Election | 18 pages, 1 figure. Published in IEEE Blockchain Technical Briefs | IEEE Blockchain Technical Briefs, (2024).
https://blockchain.ieee.org/images/files/pdf/techbriefs/tb-2024/ | null | null | cs.CR cs.DC math.NT | http://creativecommons.org/licenses/by-nc-nd/4.0/ | On June 2, 2024, Mexico held its federal elections. The majority of Mexican
citizens voted in person at the polls in this historic election. For the first
time though, Mexican citizens living outside their country were able to vote
online via a web app, either on a personal device or using an electronic voting
kiosk at one of 23 embassies and consulates in the U.S., Canada, and Europe. In
total, 144,734 people voted outside of Mexico: 122,496 on a personal device and
22,238 in-person at a kiosk. Voting was open for remote voting from 8PM, May
18, 2024 to 6PM, June 2, 2024 and was open for in-person voting from 8AM-6PM on
June 2, 2024. This article describes the technical and cryptographic tools
applied to secure the ex-patriate component of the election and to enable INE
(Mexico's National Electoral Institute) to generate provable election results
within minutes of the close of the election. This article will also describe
how the solutions we present scale to elections on a national level.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 20:46:42 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Landquist",
"Eric",
""
],
[
"Sawhney",
"Nimit",
""
],
[
"Sawhney",
"Simer",
""
]
] |
2504.13199 | Mohammad Saleh | Mohammad Saleha, Azadeh Tabatabaeib | Building Trustworthy Multimodal AI: A Review of Fairness, Transparency,
and Ethics in Vision-Language Tasks | null | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Objective: This review explores the trustworthiness of multimodal artificial
intelligence (AI) systems, specifically focusing on vision-language tasks. It
addresses critical challenges related to fairness, transparency, and ethical
implications in these systems, providing a comparative analysis of key tasks
such as Visual Question Answering (VQA), image captioning, and visual dialogue.
Background: Multimodal models, particularly vision-language models, enhance
artificial intelligence (AI) capabilities by integrating visual and textual
data, mimicking human learning processes. Despite significant advancements, the
trustworthiness of these models remains a crucial concern, particularly as AI
systems increasingly confront issues regarding fairness, transparency, and
ethics. Methods: This review examines research conducted from 2017 to 2024
focusing on forenamed core vision-language tasks. It employs a comparative
approach to analyze these tasks through the lens of trustworthiness,
underlining fairness, explainability, and ethics. This study synthesizes
findings from recent literature to identify trends, challenges, and
state-of-the-art solutions. Results: Several key findings were highlighted.
Transparency: Explainability of vision language tasks is important for user
trust. Techniques, such as attention maps and gradient-based methods, have
successfully addressed this issue. Fairness: Bias mitigation in VQA and visual
dialogue systems is essential for ensuring unbiased outcomes across diverse
demographic groups. Ethical Implications: Addressing biases in multilingual
models and ensuring ethical data handling is critical for the responsible
deployment of vision-language systems. Conclusion: This study underscores the
importance of integrating fairness, transparency, and ethical considerations in
developing vision-language models within a unified framework.
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 21:10:25 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Saleha",
"Mohammad",
""
],
[
"Tabatabaeib",
"Azadeh",
""
]
] |
2504.13200 | Mohammad Mahdi Danesh Pajouh | Mohammad Mahdi Danesh Pajouh | Efficient Brain Tumor Segmentation Using a Dual-Decoder 3D U-Net with
Attention Gates (DDUNet) | null | null | null | null | eess.IV cs.AI cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Cancer remains one of the leading causes of mortality worldwide, and among
its many forms, brain tumors are particularly notorious due to their aggressive
nature and the critical challenges involved in early diagnosis. Recent advances
in artificial intelligence have shown great promise in assisting medical
professionals with precise tumor segmentation, a key step in timely diagnosis
and treatment planning. However, many state-of-the-art segmentation methods
require extensive computational resources and prolonged training times,
limiting their practical application in resource-constrained settings. In this
work, we present a novel dual-decoder U-Net architecture enhanced with
attention-gated skip connections, designed specifically for brain tumor
segmentation from MRI scans. Our approach balances efficiency and accuracy by
achieving competitive segmentation performance while significantly reducing
training demands. Evaluated on the BraTS 2020 dataset, the proposed model
achieved Dice scores of 85.06% for Whole Tumor (WT), 80.61% for Tumor Core
(TC), and 71.26% for Enhancing Tumor (ET) in only 50 epochs, surpassing several
commonly used U-Net variants. Our model demonstrates that high-quality brain
tumor segmentation is attainable even under limited computational resources,
thereby offering a viable solution for researchers and clinicians operating
with modest hardware. This resource-efficient model has the potential to
improve early detection and diagnosis of brain tumors, ultimately contributing
to better patient outcomes
| [
{
"version": "v1",
"created": "Mon, 14 Apr 2025 22:45:33 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Pajouh",
"Mohammad Mahdi Danesh",
""
]
] |
2504.13201 | Jirui Yang | Jirui Yang, Zheyu Lin, Shuhan Yang, Zhihui Lu, Xin Du | Concept Enhancement Engineering: A Lightweight and Efficient Robust
Defense Against Jailbreak Attacks in Embodied AI | null | null | null | null | cs.CR cs.LG cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Embodied Intelligence (EI) systems integrated with large language models
(LLMs) face significant security risks, particularly from jailbreak attacks
that manipulate models into generating harmful outputs or executing unsafe
physical actions. Traditional defense strategies, such as input filtering and
output monitoring, often introduce high computational overhead or interfere
with task performance in real-time embodied scenarios. To address these
challenges, we propose Concept Enhancement Engineering (CEE), a novel defense
framework that leverages representation engineering to enhance the safety of
embodied LLMs by dynamically steering their internal activations. CEE operates
by (1) extracting multilingual safety patterns from model activations, (2)
constructing control directions based on safety-aligned concept subspaces, and
(3) applying subspace concept rotation to reinforce safe behavior during
inference. Our experiments demonstrate that CEE effectively mitigates jailbreak
attacks while maintaining task performance, outperforming existing defense
methods in both robustness and efficiency. This work contributes a scalable and
interpretable safety mechanism for embodied AI, bridging the gap between
theoretical representation engineering and practical security applications. Our
findings highlight the potential of latent-space interventions as a viable
defense paradigm against emerging adversarial threats in physically grounded AI
systems.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 03:50:04 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Jirui",
""
],
[
"Lin",
"Zheyu",
""
],
[
"Yang",
"Shuhan",
""
],
[
"Lu",
"Zhihui",
""
],
[
"Du",
"Xin",
""
]
] |
2504.13202 | Timo Laine | Timo Aukusti Laine | The Quantum LLM: Modeling Semantic Spaces with Quantum Principles | 16 pages, 6 figures | null | null | null | cs.AI cs.CL quant-ph | http://creativecommons.org/licenses/by/4.0/ | In the previous article, we presented a quantum-inspired framework for
modeling semantic representation and processing in Large Language Models
(LLMs), drawing upon mathematical tools and conceptual analogies from quantum
mechanics to offer a new perspective on these complex systems. In this paper,
we clarify the core assumptions of this model, providing a detailed exposition
of six key principles that govern semantic representation, interaction, and
dynamics within LLMs. The goal is to justify that a quantum-inspired framework
is a valid approach to studying semantic spaces. This framework offers valuable
insights into their information processing and response generation, and we
further discuss the potential of leveraging quantum computing to develop
significantly more powerful and efficient LLMs based on these principles.
| [
{
"version": "v1",
"created": "Sun, 13 Apr 2025 15:49:41 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Laine",
"Timo Aukusti",
""
]
] |
2504.13203 | Salman Rahman | Salman Rahman, Liwei Jiang, James Shiffer, Genglin Liu, Sheriff
Issaka, Md Rizwan Parvez, Hamid Palangi, Kai-Wei Chang, Yejin Choi, Saadia
Gabriel | X-Teaming: Multi-Turn Jailbreaks and Defenses with Adaptive Multi-Agents | null | null | null | null | cs.CR cs.AI cs.CL cs.LG cs.MA | http://creativecommons.org/licenses/by/4.0/ | Multi-turn interactions with language models (LMs) pose critical safety
risks, as harmful intent can be strategically spread across exchanges. Yet, the
vast majority of prior work has focused on single-turn safety, while
adaptability and diversity remain among the key challenges of multi-turn
red-teaming. To address these challenges, we present X-Teaming, a scalable
framework that systematically explores how seemingly harmless interactions
escalate into harmful outcomes and generates corresponding attack scenarios.
X-Teaming employs collaborative agents for planning, attack optimization, and
verification, achieving state-of-the-art multi-turn jailbreak effectiveness and
diversity with success rates up to 98.1% across representative leading
open-weight and closed-source models. In particular, X-Teaming achieves a 96.2%
attack success rate against the latest Claude 3.7 Sonnet model, which has been
considered nearly immune to single-turn attacks. Building on X-Teaming, we
introduce XGuard-Train, an open-source multi-turn safety training dataset that
is 20x larger than the previous best resource, comprising 30K interactive
jailbreaks, designed to enable robust multi-turn safety alignment for LMs. Our
work offers essential tools and insights for mitigating sophisticated
conversational attacks, advancing the multi-turn safety of LMs.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 16:11:28 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Rahman",
"Salman",
""
],
[
"Jiang",
"Liwei",
""
],
[
"Shiffer",
"James",
""
],
[
"Liu",
"Genglin",
""
],
[
"Issaka",
"Sheriff",
""
],
[
"Parvez",
"Md Rizwan",
""
],
[
"Palangi",
"Hamid",
""
],
[
"Chang",
"Kai-Wei",
""
],
[
"Choi",
"Yejin",
""
],
[
"Gabriel",
"Saadia",
""
]
] |
2504.13204 | Olga Grebenkova | Dmytro Kotovenko and Olga Grebenkova and Bj\"orn Ommer | EDGS: Eliminating Densification for Efficient Convergence of 3DGS | null | null | null | null | cs.GR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | 3D Gaussian Splatting reconstructs scenes by starting from a sparse
Structure-from-Motion initialization and iteratively refining
under-reconstructed regions. This process is inherently slow, as it requires
multiple densification steps where Gaussians are repeatedly split and adjusted,
following a lengthy optimization path. Moreover, this incremental approach
often leads to suboptimal renderings, particularly in high-frequency regions
where detail is critical.
We propose a fundamentally different approach: we eliminate densification
process with a one-step approximation of scene geometry using triangulated
pixels from dense image correspondences. This dense initialization allows us to
estimate rough geometry of the scene while preserving rich details from input
RGB images, providing each Gaussian with well-informed colors, scales, and
positions. As a result, we dramatically shorten the optimization path and
remove the need for densification. Unlike traditional methods that rely on
sparse keypoints, our dense initialization ensures uniform detail across the
scene, even in high-frequency regions where 3DGS and other methods struggle.
Moreover, since all splats are initialized in parallel at the start of
optimization, we eliminate the need to wait for densification to adjust new
Gaussians.
Our method not only outperforms speed-optimized models in training efficiency
but also achieves higher rendering quality than state-of-the-art approaches,
all while using only half the splats of standard 3DGS. It is fully compatible
with other 3DGS acceleration techniques, making it a versatile and efficient
solution that can be integrated with existing approaches.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 18:57:55 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kotovenko",
"Dmytro",
""
],
[
"Grebenkova",
"Olga",
""
],
[
"Ommer",
"Björn",
""
]
] |
2504.13205 | Houssam Kherraz | Houssam Kherraz | On-Device Watermarking: A Socio-Technical Imperative For Authenticity In
The Age of Generative AI | 10 pages, 3 figures, ICLR 2025,
https://openreview.net/forum?id=ygE0U21vxM | null | null | null | cs.CR cs.AI | http://creativecommons.org/licenses/by/4.0/ | As generative AI models produce increasingly realistic output, both academia
and industry are focusing on the ability to detect whether an output was
generated by an AI model or not. Many of the research efforts and policy
discourse are centered around robust watermarking of AI outputs. While plenty
of progress has been made, all watermarking and AI detection techniques face
severe limitations. In this position paper, we argue that we are adopting the
wrong approach, and should instead focus on watermarking via cryptographic
signatures trustworthy content rather than AI generated ones. For audio-visual
content, in particular, all real content is grounded in the physical world and
captured via hardware sensors. This presents a unique opportunity to watermark
at the hardware layer, and we lay out a socio-technical framework and draw
parallels with HTTPS certification and Blu-Ray verification protocols. While
acknowledging implementation challenges, we contend that hardware-based
authentication offers a more tractable path forward, particularly from a policy
perspective. As generative models approach perceptual indistinguishability, the
research community should be wary of being overly optimistic with AI
watermarking, and we argue that AI watermarking research efforts are better
spent in the text and LLM space, which are ultimately not traceable to a
physical sensor.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 20:36:52 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kherraz",
"Houssam",
""
]
] |
2504.13206 | Aniket Roy | Aniket Roy, Shubhankar Borse, Shreya Kadambi, Debasmit Das, Shweta
Mahajan, Risheek Garrepalli, Hyojin Park, Ankita Nayak, Rama Chellappa,
Munawar Hayat, Fatih Porikli | DuoLoRA : Cycle-consistent and Rank-disentangled Content-Style
Personalization | null | null | null | null | cs.GR | http://creativecommons.org/licenses/by/4.0/ | We tackle the challenge of jointly personalizing content and style from a few
examples. A promising approach is to train separate Low-Rank Adapters (LoRA)
and merge them effectively, preserving both content and style. Existing
methods, such as ZipLoRA, treat content and style as independent entities,
merging them by learning masks in LoRA's output dimensions. However, content
and style are intertwined, not independent. To address this, we propose
DuoLoRA, a content-style personalization framework featuring three key
components: (i) rank-dimension mask learning, (ii) effective merging via layer
priors, and (iii) Constyle loss, which leverages cycle-consistency in the
merging process. First, we introduce ZipRank, which performs content-style
merging within the rank dimension, offering adaptive rank flexibility and
significantly reducing the number of learnable parameters. Additionally, we
incorporate SDXL layer priors to apply implicit rank constraints informed by
each layer's content-style bias and adaptive merger initialization, enhancing
the integration of content and style. To further refine the merging process, we
introduce Constyle loss, which leverages the cycle-consistency between content
and style. Our experimental results demonstrate that DuoLoRA outperforms
state-of-the-art content-style merging methods across multiple benchmarks.
| [
{
"version": "v1",
"created": "Tue, 15 Apr 2025 23:39:28 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Roy",
"Aniket",
""
],
[
"Borse",
"Shubhankar",
""
],
[
"Kadambi",
"Shreya",
""
],
[
"Das",
"Debasmit",
""
],
[
"Mahajan",
"Shweta",
""
],
[
"Garrepalli",
"Risheek",
""
],
[
"Park",
"Hyojin",
""
],
[
"Nayak",
"Ankita",
""
],
[
"Chellappa",
"Rama",
""
],
[
"Hayat",
"Munawar",
""
],
[
"Porikli",
"Fatih",
""
]
] |
2504.13207 | Wenhua Wu | Wenhua Wu, Tong Zhao, Chensheng Peng, Lei Yang, Yintao Wei, Zhe Liu,
Hesheng Wang | BEV-GS: Feed-forward Gaussian Splatting in Bird's-Eye-View for Road
Reconstruction | null | null | null | null | cs.GR cs.RO | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Road surface is the sole contact medium for wheels or robot feet.
Reconstructing road surface is crucial for unmanned vehicles and mobile robots.
Recent studies on Neural Radiance Fields (NeRF) and Gaussian Splatting (GS)
have achieved remarkable results in scene reconstruction. However, they
typically rely on multi-view image inputs and require prolonged optimization
times. In this paper, we propose BEV-GS, a real-time single-frame road surface
reconstruction method based on feed-forward Gaussian splatting. BEV-GS consists
of a prediction module and a rendering module. The prediction module introduces
separate geometry and texture networks following Bird's-Eye-View paradigm.
Geometric and texture parameters are directly estimated from a single frame,
avoiding per-scene optimization. In the rendering module, we utilize grid
Gaussian for road surface representation and novel view synthesis, which better
aligns with road surface characteristics. Our method achieves state-of-the-art
performance on the real-world dataset RSRD. The road elevation error reduces to
1.73 cm, and the PSNR of novel view synthesis reaches 28.36 dB. The prediction
and rendering FPS is 26, and 2061, respectively, enabling high-accuracy and
real-time applications. The code will be available at:
\href{https://github.com/cat-wwh/BEV-GS}{\texttt{https://github.com/cat-wwh/BEV-GS}}
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 02:47:10 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Wu",
"Wenhua",
""
],
[
"Zhao",
"Tong",
""
],
[
"Peng",
"Chensheng",
""
],
[
"Yang",
"Lei",
""
],
[
"Wei",
"Yintao",
""
],
[
"Liu",
"Zhe",
""
],
[
"Wang",
"Hesheng",
""
]
] |
2504.13208 | Zhengyang Li | Haomin Zuo, Zhengyang Li, Jiangchuan Gong, Zhen Tian | Intelligent road crack detection and analysis based on improved YOLOv8 | Accepted by IEEE - ICAACE 2025 | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | As urbanization speeds up and traffic flow increases, the issue of pavement
distress is becoming increasingly pronounced, posing a severe threat to road
safety and service life. Traditional methods of pothole detection rely on
manual inspection, which is not only inefficient but also costly. This paper
proposes an intelligent road crack detection and analysis system, based on the
enhanced YOLOv8 deep learning framework. A target segmentation model has been
developed through the training of 4029 images, capable of efficiently and
accurately recognizing and segmenting crack regions in roads. The model also
analyzes the segmented regions to precisely calculate the maximum and minimum
widths of cracks and their exact locations. Experimental results indicate that
the incorporation of ECA and CBAM attention mechanisms substantially enhances
the model's detection accuracy and efficiency, offering a novel solution for
road maintenance and safety monitoring.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 04:50:28 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zuo",
"Haomin",
""
],
[
"Li",
"Zhengyang",
""
],
[
"Gong",
"Jiangchuan",
""
],
[
"Tian",
"Zhen",
""
]
] |
2504.13209 | Kailong Wang | Ting Bi, Chenghang Ye, Zheyu Yang, Ziyi Zhou, Cui Tang, Jun Zhang, Zui
Tao, Kailong Wang, Liting Zhou, Yang Yang, and Tianlong Yu | On the Feasibility of Using MultiModal LLMs to Execute AR Social
Engineering Attacks | null | null | null | null | cs.CR cs.AI | http://creativecommons.org/publicdomain/zero/1.0/ | Augmented Reality (AR) and Multimodal Large Language Models (LLMs) are
rapidly evolving, providing unprecedented capabilities for human-computer
interaction. However, their integration introduces a new attack surface for
social engineering. In this paper, we systematically investigate the
feasibility of orchestrating AR-driven Social Engineering attacks using
Multimodal LLM for the first time, via our proposed SEAR framework, which
operates through three key phases: (1) AR-based social context synthesis, which
fuses Multimodal inputs (visual, auditory and environmental cues); (2)
role-based Multimodal RAG (Retrieval-Augmented Generation), which dynamically
retrieves and integrates contextual data while preserving character
differentiation; and (3) ReInteract social engineering agents, which execute
adaptive multiphase attack strategies through inference interaction loops. To
verify SEAR, we conducted an IRB-approved study with 60 participants in three
experimental configurations (unassisted, AR+LLM, and full SEAR pipeline)
compiling a new dataset of 180 annotated conversations in simulated social
scenarios. Our results show that SEAR is highly effective at eliciting
high-risk behaviors (e.g., 93.3% of participants susceptible to email
phishing). The framework was particularly effective in building trust, with 85%
of targets willing to accept an attacker's call after an interaction. Also, we
identified notable limitations such as ``occasionally artificial'' due to
perceived authenticity gaps. This work provides proof-of-concept for AR-LLM
driven social engineering attacks and insights for developing defensive
countermeasures against next-generation augmented reality threats.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 05:18:36 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Bi",
"Ting",
""
],
[
"Ye",
"Chenghang",
""
],
[
"Yang",
"Zheyu",
""
],
[
"Zhou",
"Ziyi",
""
],
[
"Tang",
"Cui",
""
],
[
"Zhang",
"Jun",
""
],
[
"Tao",
"Zui",
""
],
[
"Wang",
"Kailong",
""
],
[
"Zhou",
"Liting",
""
],
[
"Yang",
"Yang",
""
],
[
"Yu",
"Tianlong",
""
]
] |
2504.13210 | Maarten Vonk | Maarten C. Vonk, Mauricio Gonzalez Soto and Anna V. Kononova | Graphical Models for Decision-Making: Integrating Causality and Game
Theory | null | null | null | null | cs.AI cs.GT cs.LG math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Causality and game theory are two influential fields that contribute
significantly to decision-making in various domains. Causality defines and
models causal relationships in complex policy problems, while game theory
provides insights into strategic interactions among stakeholders with competing
interests. Integrating these frameworks has led to significant theoretical
advancements with the potential to improve decision-making processes. However,
practical applications of these developments remain underexplored. To support
efforts toward implementation, this paper clarifies key concepts in game theory
and causality that are essential to their intersection, particularly within the
context of probabilistic graphical models. By rigorously examining these
concepts and illustrating them with intuitive, consistent examples, we clarify
the required inputs for implementing these models, provide practitioners with
insights into their application and selection across different scenarios, and
reference existing research that supports their implementation. We hope this
work encourages broader adoption of these models in real-world scenarios.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 07:25:59 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Vonk",
"Maarten C.",
""
],
[
"Soto",
"Mauricio Gonzalez",
""
],
[
"Kononova",
"Anna V.",
""
]
] |
2504.13211 | Subin Kim | Subin Kim, Hoonrae Kim, Jihyun Lee, Yejin Jeon, and Gary Geunbae Lee | Mirror: Multimodal Cognitive Reframing Therapy for Rolling with
Resistance | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent studies have explored the use of large language models (LLMs) in
psychotherapy; however, text-based cognitive behavioral therapy (CBT) models
often struggle with client resistance, which can weaken therapeutic alliance.
To address this, we propose a multimodal approach that incorporates nonverbal
cues, allowing the AI therapist to better align its responses with the client's
negative emotional state. Specifically, we introduce a new synthetic dataset,
Multimodal Interactive Rolling with Resistance (Mirror), which is a novel
synthetic dataset that pairs client statements with corresponding facial
images. Using this dataset, we train baseline Vision-Language Models (VLMs)
that can analyze facial cues, infer emotions, and generate empathetic responses
to effectively manage resistance. They are then evaluated in terms of both the
therapist's counseling skills and the strength of the therapeutic alliance in
the presence of client resistance. Our results demonstrate that Mirror
significantly enhances the AI therapist's ability to handle resistance, which
outperforms existing text-based CBT approaches.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 08:44:26 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kim",
"Subin",
""
],
[
"Kim",
"Hoonrae",
""
],
[
"Lee",
"Jihyun",
""
],
[
"Jeon",
"Yejin",
""
],
[
"Lee",
"Gary Geunbae",
""
]
] |
2504.13212 | Emmanouil Papadogiannakis | Ioannis Vlachogiannakis, Emmanouil Papadogiannakis, Panagiotis
Papadopoulos, Nicolas Kourtellis, Evangelos Markatos | I Know What You Bought Last Summer: Investigating User Data Leakage in
E-Commerce Platforms | SECRYPT 2025 - 22nd International Conference on Security and
Cryptography, 8 pages, 5 figures | null | null | null | cs.CR cs.CY | http://creativecommons.org/licenses/by/4.0/ | In the digital age, e-commerce has transformed the way consumers shop,
offering convenience and accessibility. Nevertheless, concerns about the
privacy and security of personal information shared on these platforms have
risen. In this work, we investigate user privacy violations, noting the risks
of data leakage to third-party entities. Utilizing a semi-automated data
collection approach, we examine a selection of popular online e-shops,
revealing that nearly 30% of them violate user privacy by disclosing personal
information to third parties. We unveil how minimal user interaction across
multiple e-commerce websites can result in a comprehensive privacy breach. We
observe significant data-sharing patterns with platforms like Facebook, which
use personal information to build user profiles and link them to social media
accounts.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 09:52:04 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Vlachogiannakis",
"Ioannis",
""
],
[
"Papadogiannakis",
"Emmanouil",
""
],
[
"Papadopoulos",
"Panagiotis",
""
],
[
"Kourtellis",
"Nicolas",
""
],
[
"Markatos",
"Evangelos",
""
]
] |
2504.13213 | Wei-Min Yang | Wei-Min Yang | A Model of Realistic Flavour Symmetry: Origin of Fermion Mass, Flavour
Mixing and Leptogenesis | 12 pages, 1 figure | null | null | null | physics.gen-ph hep-ph | http://creativecommons.org/licenses/by/4.0/ | I proposed a unified model of particle physic and cosmology in \cite {1},
which can successfully account for many things of the inflation, dark energy,
dark matter, neutrino mass and baryogenesis. Based on the model, I here focus
on the fermion flavour issues. I introduce a realistic flavour symmetry to
generate the fermion mass and mixing. From the model Lagrangian I derive a
neutrino mass matrix with a special structure form and fewer parameters, its
numerical solutions are precisely fitting to all the data of the neutrino mass
spectrum and lepton mixing, and clearly show a stronger predicting power. In
addition, I put forward to a new scenario of the leptogenesis, which is
implemented through the two CP-asymmetric decay modes of a superheavy neutral
Dirac-fermion, and also I establish its relationship with the neutrino mass and
mixing. This mechanism can simply and elegantly account for the baryon
asymmetry. In short, the model can address many issues of the fermion flavours,
and it is very realistic, believable and measurable instead of an illusory
theory, therefore we expect the ongoing and future experiments to test the
model.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 11:58:11 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Wei-Min",
""
]
] |
2504.13214 | Andrew Kiruluta | Andrew Kiruluta | Wavelet-based Variational Autoencoders for High-Resolution Image
Generation | null | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by/4.0/ | Variational Autoencoders (VAEs) are powerful generative models capable of
learning compact latent representations. However, conventional VAEs often
generate relatively blurry images due to their assumption of an isotropic
Gaussian latent space and constraints in capturing high-frequency details. In
this paper, we explore a novel wavelet-based approach (Wavelet-VAE) in which
the latent space is constructed using multi-scale Haar wavelet coefficients. We
propose a comprehensive method to encode the image features into multi-scale
detail and approximation coefficients and introduce a learnable noise parameter
to maintain stochasticity. We thoroughly discuss how to reformulate the
reparameterization trick, address the KL divergence term, and integrate wavelet
sparsity principles into the training objective. Our experimental evaluation on
CIFAR-10 and other high-resolution datasets demonstrates that the Wavelet-VAE
improves visual fidelity and recovers higher-resolution details compared to
conventional VAEs. We conclude with a discussion of advantages, potential
limitations, and future research directions for wavelet-based generative
modeling.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 13:51:41 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kiruluta",
"Andrew",
""
]
] |
2504.13215 | Sunia Tanweer | Sunia Tanweer, Konstantinos Mamis, Firas A. Khasawneh | Use of Topological Data Analysis for the Detection of Phenomenological
Bifurcations in Stochastic Epidemiological Models | 27 pages, 20 figures | null | null | null | q-bio.QM math.AT math.PR q-bio.PE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate predictions of stochastic compartmental models on the severity
of disease outbreaks. The models we consider are the
Susceptible-Infected-Susceptible (SIS) for bacterial infections, and the
Susceptible -Infected-Removed (SIR) for airborne diseases. Stochasticity enters
the compartmental models as random fluctuations of the contact rate, to account
for uncertainties in the disease spread. We consider three types of noise to
model the random fluctuations: the Gaussian white and Ornstein-Uhlenbeck
noises, and the logarithmic Ornstein-Uhlenbeck (logOU). The advantages of logOU
noise are its positivity and its ability to model the presence of
superspreaders. We utilize homological bifurcation plots from Topological Data
Analysis to automatically determine the shape of the long-time distributions of
the number of infected for the SIS, and removed for the SIR model, over a range
of basic reproduction numbers and relative noise intensities. LogOU noise
results in distributions that stay close to the endemic deterministic
equilibrium even for high noise intensities. For low reproduction rates and
increasing intensity, the distribution peak shifts towards zero, that is,
disease eradication, for all three noises; for logOU noise the shift is the
slowest. Our study underlines the sensitivity of model predictions to the type
of noise considered in contact rate.
| [
{
"version": "v1",
"created": "Wed, 16 Apr 2025 23:28:31 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Tanweer",
"Sunia",
""
],
[
"Mamis",
"Konstantinos",
""
],
[
"Khasawneh",
"Firas A.",
""
]
] |
2504.13216 | Seonkyu Lim | Bokwang Hwang, Seonkyu Lim, Taewoong Kim, Yongjae Geun, Sunghyun Bang,
Sohyun Park, Jihyun Park, Myeonggyu Lee, Jinwoo Lee, Yerin Kim, Jinsun Yoo,
Jingyeong Hong, Jina Park, Yongchan Kim, Suhyun Kim, Younggyun Hahm, Yiseul
Lee, Yejee Kang, Chanhyuk Yoon, Chansu Lee, Heeyewon Jeong, Jiyeon Lee,
Seonhye Gu, Hyebin Kang, Yousang Cho, Hangyeol Yoo, KyungTae Lim | KFinEval-Pilot: A Comprehensive Benchmark Suite for Korean Financial
Language Understanding | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We introduce KFinEval-Pilot, a benchmark suite specifically designed to
evaluate large language models (LLMs) in the Korean financial domain.
Addressing the limitations of existing English-centric benchmarks,
KFinEval-Pilot comprises over 1,000 curated questions across three critical
areas: financial knowledge, legal reasoning, and financial toxicity. The
benchmark is constructed through a semi-automated pipeline that combines
GPT-4-generated prompts with expert validation to ensure domain relevance and
factual accuracy. We evaluate a range of representative LLMs and observe
notable performance differences across models, with trade-offs between task
accuracy and output safety across different model families. These results
highlight persistent challenges in applying LLMs to high-stakes financial
applications, particularly in reasoning and safety. Grounded in real-world
financial use cases and aligned with the Korean regulatory and linguistic
context, KFinEval-Pilot serves as an early diagnostic tool for developing safer
and more reliable financial AI systems.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 00:12:58 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Hwang",
"Bokwang",
""
],
[
"Lim",
"Seonkyu",
""
],
[
"Kim",
"Taewoong",
""
],
[
"Geun",
"Yongjae",
""
],
[
"Bang",
"Sunghyun",
""
],
[
"Park",
"Sohyun",
""
],
[
"Park",
"Jihyun",
""
],
[
"Lee",
"Myeonggyu",
""
],
[
"Lee",
"Jinwoo",
""
],
[
"Kim",
"Yerin",
""
],
[
"Yoo",
"Jinsun",
""
],
[
"Hong",
"Jingyeong",
""
],
[
"Park",
"Jina",
""
],
[
"Kim",
"Yongchan",
""
],
[
"Kim",
"Suhyun",
""
],
[
"Hahm",
"Younggyun",
""
],
[
"Lee",
"Yiseul",
""
],
[
"Kang",
"Yejee",
""
],
[
"Yoon",
"Chanhyuk",
""
],
[
"Lee",
"Chansu",
""
],
[
"Jeong",
"Heeyewon",
""
],
[
"Lee",
"Jiyeon",
""
],
[
"Gu",
"Seonhye",
""
],
[
"Kang",
"Hyebin",
""
],
[
"Cho",
"Yousang",
""
],
[
"Yoo",
"Hangyeol",
""
],
[
"Lim",
"KyungTae",
""
]
] |
2504.13217 | Jennifer Haase | Jennifer Haase, Finn Klessascheck, Jan Mendling, Sebastian Pokutta | Sustainability via LLM Right-sizing | 17 pages, 2 Figures, 6 Tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Large language models (LLMs) have become increasingly embedded in
organizational workflows. This has raised concerns over their energy
consumption, financial costs, and data sovereignty. While performance
benchmarks often celebrate cutting-edge models, real-world deployment decisions
require a broader perspective: when is a smaller, locally deployable model
"good enough"? This study offers an empirical answer by evaluating eleven
proprietary and open-weight LLMs across ten everyday occupational tasks,
including summarizing texts, generating schedules, and drafting emails and
proposals. Using a dual-LLM-based evaluation framework, we automated task
execution and standardized evaluation across ten criteria related to output
quality, factual accuracy, and ethical responsibility. Results show that GPT-4o
delivers consistently superior performance but at a significantly higher cost
and environmental footprint. Notably, smaller models like Gemma-3 and Phi-4
achieved strong and reliable results on most tasks, suggesting their viability
in contexts requiring cost-efficiency, local deployment, or privacy. A cluster
analysis revealed three model groups -- premium all-rounders, competent
generalists, and limited but safe performers -- highlighting trade-offs between
quality, control, and sustainability. Significantly, task type influenced model
effectiveness: conceptual tasks challenged most models, while aggregation and
transformation tasks yielded better performances. We argue for a shift from
performance-maximizing benchmarks to task- and context-aware sufficiency
assessments that better reflect organizational priorities. Our approach
contributes a scalable method to evaluate AI models through a sustainability
lens and offers actionable guidance for responsible LLM deployment in practice.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 04:00:40 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Haase",
"Jennifer",
""
],
[
"Klessascheck",
"Finn",
""
],
[
"Mendling",
"Jan",
""
],
[
"Pokutta",
"Sebastian",
""
]
] |
2504.13218 | Yaguang Song | Yaguang Song, Xiaoshan Yang, Dongmei Jiang, Yaowei Wang, Changsheng Xu | Harmony: A Unified Framework for Modality Incremental Learning | null | null | null | null | cs.LG cs.AI cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Incremental learning aims to enable models to continuously acquire knowledge
from evolving data streams while preserving previously learned capabilities.
While current research predominantly focuses on unimodal incremental learning
and multimodal incremental learning where the modalities are consistent,
real-world scenarios often present data from entirely new modalities, posing
additional challenges. This paper investigates the feasibility of developing a
unified model capable of incremental learning across continuously evolving
modal sequences. To this end, we introduce a novel paradigm called Modality
Incremental Learning (MIL), where each learning stage involves data from
distinct modalities. To address this task, we propose a novel framework named
Harmony, designed to achieve modal alignment and knowledge retention, enabling
the model to reduce the modal discrepancy and learn from a sequence of distinct
modalities, ultimately completing tasks across multiple modalities within a
unified framework. Our approach introduces the adaptive compatible feature
modulation and cumulative modal bridging. Through constructing historical modal
features and performing modal knowledge accumulation and alignment, the
proposed components collaboratively bridge modal differences and maintain
knowledge retention, even with solely unimodal data available at each learning
phase.These components work in concert to establish effective modality
connections and maintain knowledge retention, even when only unimodal data is
available at each learning stage. Extensive experiments on the MIL task
demonstrate that our proposed method significantly outperforms existing
incremental learning methods, validating its effectiveness in MIL scenarios.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 06:35:01 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Song",
"Yaguang",
""
],
[
"Yang",
"Xiaoshan",
""
],
[
"Jiang",
"Dongmei",
""
],
[
"Wang",
"Yaowei",
""
],
[
"Xu",
"Changsheng",
""
]
] |
2504.13219 | Wenxuan Yang | Wenxuan Yang, Qingqu Wei, Chenxi Ma, Weimin Tan, Bo Yan | Scaling Laws for Data-Efficient Visual Transfer Learning | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Current scaling laws for visual AI models focus predominantly on large-scale
pretraining, leaving a critical gap in understanding how performance scales for
data-constrained downstream tasks. To address this limitation, this paper
establishes the first practical framework for data-efficient scaling laws in
visual transfer learning, addressing two fundamental questions: 1) How do
scaling behaviors shift when downstream tasks operate with limited data? 2)
What governs the efficacy of knowledge distillation under such constraints?
Through systematic analysis of vision tasks across data regimes (1K-1M
samples), we propose the distillation boundary theory, revealing a critical
turning point in distillation efficiency: 1) Distillation superiority: In
data-scarce conditions, distilled models significantly outperform their
non-distillation counterparts, efficiently leveraging inherited knowledge to
compensate for limited training samples. 2) Pre-training dominance: As
pre-training data increases beyond a critical threshold, non-distilled models
gradually surpass distilled versions, suggesting diminishing returns from
knowledge inheritance when sufficient task-specific data becomes available.
Empirical validation across various model scales (2.5M to 38M parameters) and
data volumes demonstrate these performance inflection points, with error
difference curves transitioning from positive to negative values at critical
data thresholds, confirming our theoretical predictions. This work redefines
scaling laws for data-limited regimes, bridging the knowledge gap between
large-scale pretraining and practical downstream adaptation, addressing a
critical barrier to understanding vision model scaling behaviors and optimizing
computational resource allocation.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 07:01:01 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Wenxuan",
""
],
[
"Wei",
"Qingqu",
""
],
[
"Ma",
"Chenxi",
""
],
[
"Tan",
"Weimin",
""
],
[
"Yan",
"Bo",
""
]
] |
2504.13220 | Md Mehedi Hasan Shawon | Ummay Maria Muna, Md. Mehedi Hasan Shawon, Md Jobayer, Sumaiya Akter,
and Saifur Rahman Sabuj | SSTAF: Spatial-Spectral-Temporal Attention Fusion Transformer for Motor
Imagery Classification | 11 pages | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Brain-computer interfaces (BCI) in electroencephalography (EEG)-based motor
imagery classification offer promising solutions in neurorehabilitation and
assistive technologies by enabling communication between the brain and external
devices. However, the non-stationary nature of EEG signals and significant
inter-subject variability cause substantial challenges for developing robust
cross-subject classification models. This paper introduces a novel
Spatial-Spectral-Temporal Attention Fusion (SSTAF) Transformer specifically
designed for upper-limb motor imagery classification. Our architecture consists
of a spectral transformer and a spatial transformer, followed by a transformer
block and a classifier network. Each module is integrated with attention
mechanisms that dynamically attend to the most discriminative patterns across
multiple domains, such as spectral frequencies, spatial electrode locations,
and temporal dynamics. The short-time Fourier transform is incorporated to
extract features in the time-frequency domain to make it easier for the model
to obtain a better feature distinction. We evaluated our SSTAF Transformer
model on two publicly available datasets, the EEGMMIDB dataset, and BCI
Competition IV-2a. SSTAF Transformer achieves an accuracy of 76.83% and 68.30%
in the data sets, respectively, outperforms traditional CNN-based architectures
and a few existing transformer-based approaches.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 07:45:14 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Muna",
"Ummay Maria",
""
],
[
"Shawon",
"Md. Mehedi Hasan",
""
],
[
"Jobayer",
"Md",
""
],
[
"Akter",
"Sumaiya",
""
],
[
"Sabuj",
"Saifur Rahman",
""
]
] |
2504.13221 | Vladimir A. Petrov | Vladimir A. Petrov | IR fixed point and low-momentum gluon propagator | null | null | null | null | hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Assuming the gluon propagator to be finite at $ \alpha_{s} /4\pi \equiv
\alpha \rightarrow \alpha_{\ast} $, where $ \alpha_{\ast} = \alpha (0) $, we
derive its low-momentum asymptotics. Self-consistency leads to some constraints
for the $ \beta $-function and the anomalous dimension of the gluon field.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 09:31:39 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Petrov",
"Vladimir A.",
""
]
] |
2504.13222 | Maurizio Maria D'Eliseo | Maurizio M. D'Eliseo | Einstein's perihelion formula and its generalization | null | Am. J. Phys. 83 (4), April 2015 | 10.1119/1.4903166 | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | Einstein's perihelion advance formula can be given a geometric interpretation
in terms of the curvature of the ellipse. The formula can be obtained by
splitting the constant term of an auxiliary polar equation for an elliptical
orbit into two parts that, when combined, lead to the expression of this
relativistic effect. Using this idea, we develop a general method for dealing
with orbital precession in the presence of central perturbing forces, and apply
the method to the determination of the total (relativistic plus Newtonian)
secular perihelion advance of the planet Mercury.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:09:19 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"D'Eliseo",
"Maurizio M.",
""
]
] |
2504.13223 | Angelos Alexopoulos Dr | Angelos Alexopoulos, Ilias Kostarakos, Christos Mylonakis, Petros
Varthalitis | The heterogeneous causal effects of the EU's Cohesion Fund | 32 pages, 10 Figures, 10 Tables | null | null | null | econ.GN econ.EM q-fin.EC | http://creativecommons.org/licenses/by/4.0/ | This paper quantifies the causal effect of cohesion policy on EU regional
output and investment focusing on one of its least studied instruments, i.e.,
the Cohesion Fund (CF). We employ modern causal inference methods to estimate
not only the local average treatment effect but also its time-varying and
heterogeneous effects across regions. Utilizing this method, we propose a novel
framework for evaluating the effectiveness of CF as an EU cohesion policy tool.
Specifically, we estimate the time varying distribution of the CF's causal
effects across EU regions and derive key distribution metrics useful for policy
evaluation. Our analysis shows that relying solely on average treatment effects
masks significant heterogeneity and can lead to misleading conclusions about
the effectiveness of the EU's cohesion policy. We find that the impact of the
CF is frontloaded, peaking within the first seven years after a region's
initial inclusion in the program. The distribution of the effects during this
first seven-year cycle of funding is right skewed with relatively thick tails.
This indicates positive effects but unevenly distributed across regions.
Moreover, the magnitude of the CF effect is inversely related to a region's
relative position in the initial distribution of output, i.e., relatively
poorer recipient regions experience higher effects compared to relatively
richer regions. Finally, we find a non-linear relationship with diminishing
returns, whereby the impact of CF declines as the ratio of CF funds received to
a region's gross value added (GVA) increases.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:22:54 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Alexopoulos",
"Angelos",
""
],
[
"Kostarakos",
"Ilias",
""
],
[
"Mylonakis",
"Christos",
""
],
[
"Varthalitis",
"Petros",
""
]
] |
2504.13224 | Fuwei Liu | Fuwei Liu | ICAS: IP Adapter and ControlNet-based Attention Structure for
Multi-Subject Style Transfer Optimization | 10 pages, 6 figures | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generating multi-subject stylized images remains a significant challenge due
to the ambiguity in defining style attributes (e.g., color, texture,
atmosphere, and structure) and the difficulty in consistently applying them
across multiple subjects. Although recent diffusion-based text-to-image models
have achieved remarkable progress, existing methods typically rely on
computationally expensive inversion procedures or large-scale stylized
datasets. Moreover, these methods often struggle with maintaining multi-subject
semantic fidelity and are limited by high inference costs. To address these
limitations, we propose ICAS (IP-Adapter and ControlNet-based Attention
Structure), a novel framework for efficient and controllable multi-subject
style transfer. Instead of full-model tuning, ICAS adaptively fine-tunes only
the content injection branch of a pre-trained diffusion model, thereby
preserving identity-specific semantics while enhancing style controllability.
By combining IP-Adapter for adaptive style injection with ControlNet for
structural conditioning, our framework ensures faithful global layout
preservation alongside accurate local style synthesis. Furthermore, ICAS
introduces a cyclic multi-subject content embedding mechanism, which enables
effective style transfer under limited-data settings without the need for
extensive stylized corpora. Extensive experiments show that ICAS achieves
superior performance in structure preservation, style consistency, and
inference efficiency, establishing a new paradigm for multi-subject style
transfer in real-world applications.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 10:48:11 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Liu",
"Fuwei",
""
]
] |
2504.13225 | Jose Gamboa Manuel | E. Baro, J.F. Fernando and J.M. Gamboa | Intermediate algebras of semialgebraic functions | null | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We characterize intermediate $\mathbb{R}$-algebras $A$ between the ring of
semialgebraic functions ${\mathcal S}(X)$ and the ring ${\mathcal S}^*(X)$ of
bounded semialgebraic functions on a semialgebraic set $X$ as rings of
fractions of ${\mathcal S}(X)$. This allows us to compute the Krull dimension
of $A$, the transcendence degree over $\mathbb{R}$ of the residue fields of $A$
and to obtain a \L ojasiewicz inequality and a Nullstellensatz for archimedean
$\mathbb{R}$-algebras $A$. In addition we study intermediate
$\mathbb{R}$-algebras generated by proper ideals and we prove an extension
theorem for functions in such $\mathbb{R}$-algebras.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 11:49:47 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Baro",
"E.",
""
],
[
"Fernando",
"J. F.",
""
],
[
"Gamboa",
"J. M.",
""
]
] |
2504.13226 | Jia Wang | Jia Wang, Jie Hu, Xiaoqi Ma, Hanghang Ma, Xiaoming Wei and Enhua Wu | Image Editing with Diffusion Models: A Survey | null | null | null | null | cs.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With deeper exploration of diffusion model, developments in the field of
image generation have triggered a boom in image creation. As the quality of
base-model generated images continues to improve, so does the demand for
further application like image editing. In recent years, many remarkable works
are realizing a wide variety of editing effects. However, the wide variety of
editing types and diverse editing approaches have made it difficult for
researchers to establish a comprehensive view of the development of this field.
In this survey, we summarize the image editing field from four aspects: tasks
definition, methods classification, results evaluation and editing datasets.
First, we provide a definition of image editing, which in turn leads to a
variety of editing task forms from the perspective of operation parts and
manipulation actions. Subsequently, we categorize and summary methods for
implementing editing into three categories: inversion-based, fine-tuning-based
and adapter-based. In addition, we organize the currently used metrics,
available datasets and corresponding construction methods. At the end, we
present some visions for the future development of the image editing field
based on the previous summaries.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 11:57:41 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Wang",
"Jia",
""
],
[
"Hu",
"Jie",
""
],
[
"Ma",
"Xiaoqi",
""
],
[
"Ma",
"Hanghang",
""
],
[
"Wei",
"Xiaoming",
""
],
[
"Wu",
"Enhua",
""
]
] |
2504.13227 | Weijie Shi | Weijie Shi, Jipeng Zhang, Yaguang Wu, Jingzhi Fang, Ruiyuan Zhang,
Jiajie Xu, Jia Zhu, Hao Chen, Yao Zhao, Sirui Han, Xiaofang Zhou | DIDS: Domain Impact-aware Data Sampling for Large Language Model
Training | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Large language models (LLMs) are commonly trained on multi-domain datasets,
where domain sampling strategies significantly impact model performance due to
varying domain importance across downstream tasks. Existing approaches for
optimizing domain-level sampling strategies struggle with maintaining
intra-domain consistency and accurately measuring domain impact. In this paper,
we present Domain Impact-aware Data Sampling (DIDS). To ensure intra-domain
consistency, a gradient clustering algorithm is proposed to group training data
based on their learning effects, where a proxy language model and
dimensionality reduction are employed to reduce computational overhead. To
accurately measure domain impact, we develop a Fisher Information Matrix (FIM)
guided metric that quantifies how domain-specific parameter updates affect the
model's output distributions on downstream tasks, with theoretical guarantees.
Furthermore, to determine optimal sampling ratios, DIDS combines both the
FIM-guided domain impact assessment and loss learning trajectories that
indicate domain-specific potential, while accounting for diminishing marginal
returns. Extensive experiments demonstrate that DIDS achieves 3.4% higher
average performance while maintaining comparable training efficiency.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:09:38 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Shi",
"Weijie",
""
],
[
"Zhang",
"Jipeng",
""
],
[
"Wu",
"Yaguang",
""
],
[
"Fang",
"Jingzhi",
""
],
[
"Zhang",
"Ruiyuan",
""
],
[
"Xu",
"Jiajie",
""
],
[
"Zhu",
"Jia",
""
],
[
"Chen",
"Hao",
""
],
[
"Zhao",
"Yao",
""
],
[
"Han",
"Sirui",
""
],
[
"Zhou",
"Xiaofang",
""
]
] |
2504.13228 | Chiara Th\"oni | Anna C.M. Th\"oni, Yoram Bachrach and Tal Kachman | Modelling Mean-Field Games with Neural Ordinary Differential Equations | null | null | null | null | cs.LG cs.GT | http://creativecommons.org/licenses/by/4.0/ | Mean-field game theory relies on approximating games that would otherwise
have been intractable to model. While the games can be solved analytically via
the associated system of partial derivatives, this approach is not model-free,
can lead to the loss of the existence or uniqueness of solutions and may suffer
from modelling bias. To reduce the dependency between the model and the game,
we combine mean-field game theory with deep learning in the form of neural
ordinary differential equations. The resulting model is data-driven,
lightweight and can learn extensive strategic interactions that are hard to
capture using mean-field theory alone. In addition, the model is based on
automatic differentiation, making it more robust and objective than approaches
based on finite differences. We highlight the efficiency and flexibility of our
approach by solving three mean-field games that vary in their complexity,
observability and the presence of noise. Using these results, we show that the
model is flexible, lightweight and requires few observations to learn the
distribution underlying the data.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:38:42 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Thöni",
"Anna C. M.",
""
],
[
"Bachrach",
"Yoram",
""
],
[
"Kachman",
"Tal",
""
]
] |
2504.13229 | Yifei Wang | Yifei Wang, Qi Liu, Fuli Min, and Honghao Wang | PSG-MAE: Robust Multitask Sleep Event Monitoring using Multichannel PSG
Reconstruction and Inter-channel Contrastive Learning | 11 pages, 5 figures | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Polysomnography (PSG) signals are essential for studying sleep processes and
diagnosing sleep disorders. Analyzing PSG data through deep neural networks
(DNNs) for automated sleep monitoring has become increasingly feasible.
However, the limited availability of datasets for certain sleep events often
leads to DNNs focusing on a single task with a single-sourced training dataset.
As a result, these models struggle to transfer to new sleep events and lack
robustness when applied to new datasets. To address these challenges, we
propose PSG-MAE, a mask autoencoder (MAE) based pre-training framework. By
performing self-supervised learning on a large volume of unlabeled PSG data,
PSG-MAE develops a robust feature extraction network that can be broadly
applied to various sleep event monitoring tasks. Unlike conventional MAEs,
PSG-MAE generates complementary masks across PSG channels, integrates a
multichannel signal reconstruction method, and employs a self-supervised
inter-channel contrastive learning (ICCL) strategy. This approach enables the
encoder to capture temporal features from each channel while simultaneously
learning latent relationships between channels, thereby enhancing the
utilization of multichannel information. Experimental results show that PSG-MAE
effectively captures both temporal details and inter-channel information from
PSG signals. When the encoder pre-trained through PSG-MAE is fine-tuned with
downstream feature decomposition networks, it achieves an accuracy of 83.7% for
sleep staging and 90.45% for detecting obstructive sleep apnea, which
highlights the framework's robustness and broad applicability.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:43:16 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Wang",
"Yifei",
""
],
[
"Liu",
"Qi",
""
],
[
"Min",
"Fuli",
""
],
[
"Wang",
"Honghao",
""
]
] |
2504.13230 | Norbert Hegyv\'ari | Norbert Hegyv\'ari | Note on the sumset of squares | 5 pages, comments are welcome | null | null | null | math.CO | http://creativecommons.org/publicdomain/zero/1.0/ | It is proved that for any non-empty finite subset $Q$ of the square numbers,
$ |Q+Q|\geq C'|Q|(\log |Q|)^{1/3+o(1)} $.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 13:45:59 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Hegyvári",
"Norbert",
""
]
] |
2504.13231 | Isar Nejadgholi | Braeden Sherritt and Isar Nejadgholi and Marzieh Amini | WildFireCan-MMD: A Multimodal dataset for Classification of
User-generated Content During Wildfires in Canada | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Rapid information access is vital during wildfires, yet traditional data
sources are slow and costly. Social media offers real-time updates, but
extracting relevant insights remains a challenge. We present WildFireCan-MMD, a
new multimodal dataset of X posts from recent Canadian wildfires, annotated
across 13 key themes. Evaluating both Vision Language Models and custom-trained
classifiers, we show that while zero-shot prompting offers quick deployment,
even simple trained models outperform them when labelled data is available, by
up to 23%. Our findings highlight the enduring importance of tailored datasets
and task-specific training. Importantly, such datasets should be localized, as
disaster response requirements vary across regions and contexts.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:43:56 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Sherritt",
"Braeden",
""
],
[
"Nejadgholi",
"Isar",
""
],
[
"Amini",
"Marzieh",
""
]
] |
2504.13232 | Sayed Pouria Talebi | Sayed Pouria Talebi, Clive Cheong Took, and Danilo P. Mandic | A Quantum of Learning: Using Quaternion Algebra to Model Learning on
Quantum Devices | null | null | null | null | quant-ph cs.LG math.QA stat.ML | http://creativecommons.org/licenses/by/4.0/ | This article considers the problem of designing adaption and optimisation
techniques for training quantum learning machines. To this end, the division
algebra of quaternions is used to derive an effective model for representing
computation and measurement operations on qubits. In turn, the derived model,
serves as the foundation for formulating an adaptive learning problem on
principal quantum learning units, thereby establishing quantum information
processing units akin to that of neurons in classical approaches. Then,
leveraging the modern HR-calculus, a comprehensive training framework for
learning on quantum machines is developed. The quaternion-valued model
accommodates mathematical tractability and establishment of performance
criteria, such as convergence conditions.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 14:51:21 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Talebi",
"Sayed Pouria",
""
],
[
"Took",
"Clive Cheong",
""
],
[
"Mandic",
"Danilo P.",
""
]
] |
2504.13233 | Alireza Rafiei | Alireza Rafiei, Gari D. Clifford, Nasim Katebi | Auto-FEDUS: Autoregressive Generative Modeling of Doppler Ultrasound
Signals from Fetal Electrocardiograms | AAAI 2025 Workshop on Large Language Models and Generative AI for
Health | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fetal health monitoring through one-dimensional Doppler ultrasound (DUS)
signals offers a cost-effective and accessible approach that is increasingly
gaining interest. Despite its potential, the development of machine learning
based techniques to assess the health condition of mothers and fetuses using
DUS signals remains limited. This scarcity is primarily due to the lack of
extensive DUS datasets with a reliable reference for interpretation and data
imbalance across different gestational ages. In response, we introduce a novel
autoregressive generative model designed to map fetal electrocardiogram (FECG)
signals to corresponding DUS waveforms (Auto-FEDUS). By leveraging a neural
temporal network based on dilated causal convolutions that operate directly on
the waveform level, the model effectively captures both short and long-range
dependencies within the signals, preserving the integrity of generated data.
Cross-subject experiments demonstrate that Auto-FEDUS outperforms conventional
generative architectures across both time and frequency domain evaluations,
producing DUS signals that closely resemble the morphology of their real
counterparts. The realism of these synthesized signals was further gauged using
a quality assessment model, which classified all as good quality, and a heart
rate estimation model, which produced comparable results for generated and real
data, with a Bland-Altman limit of 4.5 beats per minute. This advancement
offers a promising solution for mitigating limited data availability and
enhancing the training of DUS-based fetal models, making them more effective
and generalizable.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 15:25:52 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Rafiei",
"Alireza",
""
],
[
"Clifford",
"Gari D.",
""
],
[
"Katebi",
"Nasim",
""
]
] |
2504.13234 | Hanyu Zhang | Hanyu Zhang, Zhen Xing, Wenxuan Yang, Chenxi Ma, Weimin Tan, Bo Yan | Non-Uniform Class-Wise Coreset Selection: Characterizing Category
Difficulty for Data-Efficient Transfer Learning | 11pages | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As transfer learning models and datasets grow larger, efficient adaptation
and storage optimization have become critical needs. Coreset selection
addresses these challenges by identifying and retaining the most informative
samples, constructing a compact subset for target domain training. However,
current methods primarily rely on instance-level difficulty assessments,
overlooking crucial category-level characteristics and consequently
under-representing minority classes. To overcome this limitation, we propose
Non-Uniform Class-Wise Coreset Selection (NUCS), a novel framework that
integrates both class-level and instance-level criteria. NUCS automatically
allocates data selection budgets for each class based on intrinsic category
difficulty and adaptively selects samples within optimal difficulty ranges. By
explicitly incorporating category-specific insights, our approach achieves a
more balanced and representative coreset, addressing key shortcomings of prior
methods. Comprehensive theoretical analysis validates the rationale behind
adaptive budget allocation and sample selection, while extensive experiments
across 14 diverse datasets and model architectures demonstrate NUCS's
consistent improvements over state-of-the-art methods, achieving superior
accuracy and computational efficiency. Notably, on CIFAR100 and Food101, NUCS
matches full-data training accuracy while retaining just 30% of samples and
reducing computation time by 60%. Our work highlights the importance of
characterizing category difficulty in coreset selection, offering a robust and
data-efficient solution for transfer learning.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 15:40:51 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zhang",
"Hanyu",
""
],
[
"Xing",
"Zhen",
""
],
[
"Yang",
"Wenxuan",
""
],
[
"Ma",
"Chenxi",
""
],
[
"Tan",
"Weimin",
""
],
[
"Yan",
"Bo",
""
]
] |
2504.13235 | Weijian Liu | Daipeng Xiao, Weijian Liu, Jun Liu, Yuntao Wu, Qinglei Du and
Xiaoqiang Hua | Bayesian Rao test for distributed target detection in interference and
noise with limited training data | 14 pages,18 figures | null | null | null | stat.ME cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | This paper has studied the problem of detecting a range-spread target in
interference and noise when the number of training data is limited. The
interference is located within a certain subspace with an unknown coordinate,
while the noise follows a Gaussian distribution with an unknown covariance
matrix. We concentrate on the scenarios where the training data are limited and
employ a Bayesian framework to ffnd a solution. Speciffcally, the covariance
matrix is assumed to follow an inverse Wishart distribution. Then, we introduce
the Bayesian detector according to the Rao test, which, demonstrated by both
simulation experiment and real data, has superior detection performance to the
existing detectors in certain situations.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:02:48 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Xiao",
"Daipeng",
""
],
[
"Liu",
"Weijian",
""
],
[
"Liu",
"Jun",
""
],
[
"Wu",
"Yuntao",
""
],
[
"Du",
"Qinglei",
""
],
[
"Hua",
"Xiaoqiang",
""
]
] |
2504.13236 | Aleksandr Katrutsa | Aleksandr Mikhalev, Aleksandr Katrutsa, Konstantin Sozykin, Ivan
Oseledets | NNTile: a machine learning framework capable of training extremely large
GPT language models on a single node | null | null | null | null | cs.LG cs.MS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This study presents an NNTile framework for training large deep neural
networks in heterogeneous clusters. The NNTile is based on a StarPU library,
which implements task-based parallelism and schedules all provided tasks onto
all available processing units (CPUs and GPUs). It means that a particular
operation, necessary to train a large neural network, can be performed on any
of the CPU cores or GPU devices, depending on automatic scheduling decisions.
Such an approach shifts the burden of deciding where to compute and when to
communicate from a human being to an automatic decision maker, whether a simple
greedy heuristic or a complex AI-based software. The performance of the
presented tool for training large language models is demonstrated in extensive
numerical experiments.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:22:32 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Mikhalev",
"Aleksandr",
""
],
[
"Katrutsa",
"Aleksandr",
""
],
[
"Sozykin",
"Konstantin",
""
],
[
"Oseledets",
"Ivan",
""
]
] |
2504.13237 | Yan Yang | Yan Yang, Yixia Li, Hongru Wang, Xuetao Wei, Jianqiao Yu, Yun Chen,
Guanhua Chen | ImPart: Importance-Aware Delta-Sparsification for Improved Model
Compression and Merging in LLMs | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the proliferation of task-specific large language models, delta
compression has emerged as a method to mitigate the resource challenges of
deploying numerous such models by effectively compressing the delta model
parameters. Previous delta-sparsification methods either remove parameters
randomly or truncate singular vectors directly after singular value
decomposition (SVD). However, these methods either disregard parameter
importance entirely or evaluate it with too coarse a granularity. In this work,
we introduce ImPart, a novel importance-aware delta sparsification approach.
Leveraging SVD, it dynamically adjusts sparsity ratios of different singular
vectors based on their importance, effectively retaining crucial task-specific
knowledge even at high sparsity ratios. Experiments show that ImPart achieves
state-of-the-art delta sparsification performance, demonstrating $2\times$
higher compression ratio than baselines at the same performance level. When
integrated with existing methods, ImPart sets a new state-of-the-art on delta
quantization and model merging.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:39:36 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Yang",
"Yan",
""
],
[
"Li",
"Yixia",
""
],
[
"Wang",
"Hongru",
""
],
[
"Wei",
"Xuetao",
""
],
[
"Yu",
"Jianqiao",
""
],
[
"Chen",
"Yun",
""
],
[
"Chen",
"Guanhua",
""
]
] |
2504.13238 | David Kipping | David Kipping | Exoplaneteers Keep Calling Plots "Allan Variance" Plots When They Aren't | Community commentary article, not intended for journal submission | null | null | null | astro-ph.IM astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | I highlight that there is a substantial number of papers (at least 11
published since 2024) which all refer to a specific type of plot as an "Allan
variance" plot, when in fact they seem to be plotting the standard deviation of
the residuals versus bin size. The Allan variance quantifies the stability of a
time series by calculating the average squared difference between successive
time-averaged segments over a specified interval; it is not equivalent to the
standard deviation. This misattribution seems particularly prolific in the
exoplanet transit spectroscopy community. However, I emphasize that it does not
impact the scientific analyses presented in those works. I discuss where this
confusion seems to stem from and encourage the community to ensure statistical
measures are named correctly to avoid confusion.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:49:32 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Kipping",
"David",
""
]
] |
2504.13239 | Maurizio Maria D'Eliseo | Maurizio M. D'Eliseo | The first-order orbital equation | American Journal of Physics, 2007 | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | We derive the first-order orbital equation employing a complex variable
formalism. We then examine Newton's theorem on precessing orbits and apply it
to the perihelion shift of an elliptic orbit in general relativity. It is found
that corrections to the inverse-square gravitational force law formally similar
to that required by general relativity were suggested by Clairaut in the 18th
century.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 16:54:30 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"D'Eliseo",
"Maurizio M.",
""
]
] |
2504.13240 | Andrey Antipov | Morteza Aghaee, Zulfi Alam, Mariusz Andrzejczuk, Andrey E. Antipov,
Mikhail Astafev, Amin Barzegar, Bela Bauer, Jonathan Becker, Umesh Kumar
Bhaskar, Alex Bocharov, Srini Boddapati, David Bohn, Jouri Bommer, Leo
Bourdet, Samuel Boutin, Benjamin J. Chapman, Sohail Chatoor, Anna Wulff
Christensen, Patrick Codd, William S. Cole, Paul Cooper, Fabiano Corsetti,
Ajuan Cui, Andreas Ekefj\"ard, Saeed Fallahi, Luca Galletti, Geoff Gardner,
Deshan Govender, Flavio Griggio, Ruben Grigoryan, Sebastian Grijalva, Sergei
Gronin, Jan Gukelberger, Marzie Hamdast, Esben Bork Hansen, Sebastian Heedt,
Samantha Ho, Laurens Holgaard, Kevin Van Hoogdalem, Jinnapat Indrapiromkul,
Henrik Ingerslev, Lovro Ivancevic, Thomas Jensen, Jaspreet Jhoja, Jeffrey
Jones, Konstantin V. Kalashnikov, Ray Kallaher, Rachpon Kalra, Farhad Karimi,
Torsten Karzig, Maren Elisabeth Kloster, Christina Knapp, Jonne Koski, Pasi
Kostamo, Tom Laeven, Gijs de Lange, Thorvald Larsen, Jason Lee, Kyunghoon
Lee, Grant Leum, Kongyi Li, Tyler Lindemann, Matthew Looij, Marijn Lucas,
Roman Lutchyn, Morten Hannibal Madsen, Nash Madulid, Michael Manfra, Signe
Brynold Markussen, Esteban Martinez, Marco Mattila, Robert McNeil, Ryan V.
Mishmash, Gopakumar Mohandas, Christian Mollgaard, Michiel de Moor, Trevor
Morgan, George Moussa, Chetan Nayak, William Hvidtfelt Padk{\ae}r Nielsen,
Jens Hedegaard Nielsen, Mike Nystrom, Eoin O'Farrell, Keita Otani, Karl
Petersson, Luca Petit, Dima Pikulin, Mohana Rajpalke, Alejandro Alcaraz
Ramirez, Katrine Rasmussen, David Razmadze, Yuan Ren, Ken Reneris, Ivan A.
Sadovskyy, Lauri Sainiemi, Juan Carlos Estrada Salda\~na, Irene Sanlorenzo,
Emma Schmidgall, Cristina Sfiligoj, Sarat Sinha, Thomas Soerensen, Patrick
Sohr, Toma\v{s} Stankevi\v{c}, Lieuwe Stek, Eric Stuppard, Henri Suominen,
Judith Suter, Sam Teicher, Nivetha Thiyagarajah, Raj Tholapi, Mason Thomas,
Emily Toomey, Josh Tracy, Michelle Turley, Shivendra Upadhyay, Ivan Urban,
Dmitrii V. Viazmitinov, Dominik Vogel, John Watson, Alex Webster, Joseph
Weston, Georg W. Winkler, David J. Van Woerkom, Brian Paquelet W\"utz, Chung
Kai Yang, Emrah Yucelen, Jes\'us Herranz Zamorano, Roland Zeisel, Guoji
Zheng, Justin Zilke | Response to recent comments on Phys. Rev. B 107, 245423 (2023) and
Subsection S4.3 of the Supp. Info. for Nature 638, 651-655 (2025) | Response to arXiv:2502.19560 and arXiv:2503.08944. 11 pages, 5
figures, 2 tables, code for reproduction | null | null | null | cond-mat.mes-hall quant-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The topological gap protocol (TGP) is a statistical test designed to identify
a topological phase with high confidence and without human bias. It is used to
determine a promising parameter regime for operating topological qubits. The
protocol's key metric is the probability of incorrectly identifying a trivial
region as topological, referred to as the false discovery rate (FDR). Two
recent manuscripts [arXiv:2502.19560, arXiv:2503.08944] engage with the
topological gap protocol and its use in Phys. Rev. B 107, 245423 (2023) and
Subsection S4.3 of the Supplementary Information for Nature 638, 651-655
(2025), although they do not explicitly dispute the main results of either one.
We demonstrate that the objections in arXiv:2502.19560 and arXiv:2503.08944 are
unfounded, and we uphold the conclusions of Phys. Rev. B 107, 245423 (2023) and
Nature 638, 651-655 (2025). Specifically, we show that no flaws have been
identified in our estimate of the false discovery rate (FDR). We provide a
point-by-point rebuttal of the comments in arXiv:2502.19560 and
arXiv:2503.08944.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:14:32 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Aghaee",
"Morteza",
""
],
[
"Alam",
"Zulfi",
""
],
[
"Andrzejczuk",
"Mariusz",
""
],
[
"Antipov",
"Andrey E.",
""
],
[
"Astafev",
"Mikhail",
""
],
[
"Barzegar",
"Amin",
""
],
[
"Bauer",
"Bela",
""
],
[
"Becker",
"Jonathan",
""
],
[
"Bhaskar",
"Umesh Kumar",
""
],
[
"Bocharov",
"Alex",
""
],
[
"Boddapati",
"Srini",
""
],
[
"Bohn",
"David",
""
],
[
"Bommer",
"Jouri",
""
],
[
"Bourdet",
"Leo",
""
],
[
"Boutin",
"Samuel",
""
],
[
"Chapman",
"Benjamin J.",
""
],
[
"Chatoor",
"Sohail",
""
],
[
"Christensen",
"Anna Wulff",
""
],
[
"Codd",
"Patrick",
""
],
[
"Cole",
"William S.",
""
],
[
"Cooper",
"Paul",
""
],
[
"Corsetti",
"Fabiano",
""
],
[
"Cui",
"Ajuan",
""
],
[
"Ekefjärd",
"Andreas",
""
],
[
"Fallahi",
"Saeed",
""
],
[
"Galletti",
"Luca",
""
],
[
"Gardner",
"Geoff",
""
],
[
"Govender",
"Deshan",
""
],
[
"Griggio",
"Flavio",
""
],
[
"Grigoryan",
"Ruben",
""
],
[
"Grijalva",
"Sebastian",
""
],
[
"Gronin",
"Sergei",
""
],
[
"Gukelberger",
"Jan",
""
],
[
"Hamdast",
"Marzie",
""
],
[
"Hansen",
"Esben Bork",
""
],
[
"Heedt",
"Sebastian",
""
],
[
"Ho",
"Samantha",
""
],
[
"Holgaard",
"Laurens",
""
],
[
"Van Hoogdalem",
"Kevin",
""
],
[
"Indrapiromkul",
"Jinnapat",
""
],
[
"Ingerslev",
"Henrik",
""
],
[
"Ivancevic",
"Lovro",
""
],
[
"Jensen",
"Thomas",
""
],
[
"Jhoja",
"Jaspreet",
""
],
[
"Jones",
"Jeffrey",
""
],
[
"Kalashnikov",
"Konstantin V.",
""
],
[
"Kallaher",
"Ray",
""
],
[
"Kalra",
"Rachpon",
""
],
[
"Karimi",
"Farhad",
""
],
[
"Karzig",
"Torsten",
""
],
[
"Kloster",
"Maren Elisabeth",
""
],
[
"Knapp",
"Christina",
""
],
[
"Koski",
"Jonne",
""
],
[
"Kostamo",
"Pasi",
""
],
[
"Laeven",
"Tom",
""
],
[
"de Lange",
"Gijs",
""
],
[
"Larsen",
"Thorvald",
""
],
[
"Lee",
"Jason",
""
],
[
"Lee",
"Kyunghoon",
""
],
[
"Leum",
"Grant",
""
],
[
"Li",
"Kongyi",
""
],
[
"Lindemann",
"Tyler",
""
],
[
"Looij",
"Matthew",
""
],
[
"Lucas",
"Marijn",
""
],
[
"Lutchyn",
"Roman",
""
],
[
"Madsen",
"Morten Hannibal",
""
],
[
"Madulid",
"Nash",
""
],
[
"Manfra",
"Michael",
""
],
[
"Markussen",
"Signe Brynold",
""
],
[
"Martinez",
"Esteban",
""
],
[
"Mattila",
"Marco",
""
],
[
"McNeil",
"Robert",
""
],
[
"Mishmash",
"Ryan V.",
""
],
[
"Mohandas",
"Gopakumar",
""
],
[
"Mollgaard",
"Christian",
""
],
[
"de Moor",
"Michiel",
""
],
[
"Morgan",
"Trevor",
""
],
[
"Moussa",
"George",
""
],
[
"Nayak",
"Chetan",
""
],
[
"Nielsen",
"William Hvidtfelt Padkær",
""
],
[
"Nielsen",
"Jens Hedegaard",
""
],
[
"Nystrom",
"Mike",
""
],
[
"O'Farrell",
"Eoin",
""
],
[
"Otani",
"Keita",
""
],
[
"Petersson",
"Karl",
""
],
[
"Petit",
"Luca",
""
],
[
"Pikulin",
"Dima",
""
],
[
"Rajpalke",
"Mohana",
""
],
[
"Ramirez",
"Alejandro Alcaraz",
""
],
[
"Rasmussen",
"Katrine",
""
],
[
"Razmadze",
"David",
""
],
[
"Ren",
"Yuan",
""
],
[
"Reneris",
"Ken",
""
],
[
"Sadovskyy",
"Ivan A.",
""
],
[
"Sainiemi",
"Lauri",
""
],
[
"Saldaña",
"Juan Carlos Estrada",
""
],
[
"Sanlorenzo",
"Irene",
""
],
[
"Schmidgall",
"Emma",
""
],
[
"Sfiligoj",
"Cristina",
""
],
[
"Sinha",
"Sarat",
""
],
[
"Soerensen",
"Thomas",
""
],
[
"Sohr",
"Patrick",
""
],
[
"Stankevič",
"Tomaš",
""
],
[
"Stek",
"Lieuwe",
""
],
[
"Stuppard",
"Eric",
""
],
[
"Suominen",
"Henri",
""
],
[
"Suter",
"Judith",
""
],
[
"Teicher",
"Sam",
""
],
[
"Thiyagarajah",
"Nivetha",
""
],
[
"Tholapi",
"Raj",
""
],
[
"Thomas",
"Mason",
""
],
[
"Toomey",
"Emily",
""
],
[
"Tracy",
"Josh",
""
],
[
"Turley",
"Michelle",
""
],
[
"Upadhyay",
"Shivendra",
""
],
[
"Urban",
"Ivan",
""
],
[
"Viazmitinov",
"Dmitrii V.",
""
],
[
"Vogel",
"Dominik",
""
],
[
"Watson",
"John",
""
],
[
"Webster",
"Alex",
""
],
[
"Weston",
"Joseph",
""
],
[
"Winkler",
"Georg W.",
""
],
[
"Van Woerkom",
"David J.",
""
],
[
"Wütz",
"Brian Paquelet",
""
],
[
"Yang",
"Chung Kai",
""
],
[
"Yucelen",
"Emrah",
""
],
[
"Zamorano",
"Jesús Herranz",
""
],
[
"Zeisel",
"Roland",
""
],
[
"Zheng",
"Guoji",
""
],
[
"Zilke",
"Justin",
""
]
] |
2504.13241 | Paul Ghanem | Paul Ghanem, Michael Potter, Owen Howell, Pau Closas, Alireza
Ramezani, Deniz Erdogmus, Robert Platt, Tales Imbiriba | Recursive Deep Inverse Reinforcement Learning | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Inferring an adversary's goals from exhibited behavior is crucial for
counterplanning and non-cooperative multi-agent systems in domains like
cybersecurity, military, and strategy games. Deep Inverse Reinforcement
Learning (IRL) methods based on maximum entropy principles show promise in
recovering adversaries' goals but are typically offline, require large batch
sizes with gradient descent, and rely on first-order updates, limiting their
applicability in real-time scenarios. We propose an online Recursive Deep
Inverse Reinforcement Learning (RDIRL) approach to recover the cost function
governing the adversary actions and goals. Specifically, we minimize an upper
bound on the standard Guided Cost Learning (GCL) objective using sequential
second-order Newton updates, akin to the Extended Kalman Filter (EKF), leading
to a fast (in terms of convergence) learning algorithm. We demonstrate that
RDIRL is able to recover cost and reward functions of expert agents in standard
and adversarial benchmark tasks. Experiments on benchmark tasks show that our
proposed approach outperforms several leading IRL algorithms.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:39:35 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ghanem",
"Paul",
""
],
[
"Potter",
"Michael",
""
],
[
"Howell",
"Owen",
""
],
[
"Closas",
"Pau",
""
],
[
"Ramezani",
"Alireza",
""
],
[
"Erdogmus",
"Deniz",
""
],
[
"Platt",
"Robert",
""
],
[
"Imbiriba",
"Tales",
""
]
] |
2504.13242 | Muhammad Ahmad | Muhammad Ahmad, Manuel Mazzara, Salvatore Distefano, Adil Mehmood Khan | Dynamic Memory-enhanced Transformer for Hyperspectral Image
Classification | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Hyperspectral image (HSI) classification remains a challenging task due to
the intricate spatial-spectral correlations. Existing transformer models excel
in capturing long-range dependencies but often suffer from information
redundancy and attention inefficiencies, limiting their ability to model
fine-grained relationships crucial for HSI classification. To overcome these
limitations, this work proposes MemFormer, a lightweight and memory-enhanced
transformer. MemFormer introduces a memory-enhanced multi-head attention
mechanism that iteratively refines a dynamic memory module, enhancing feature
extraction while reducing redundancy across layers. Additionally, a dynamic
memory enrichment strategy progressively captures complex spatial and spectral
dependencies, leading to more expressive feature representations. To further
improve structural consistency, we incorporate a spatial-spectral positional
encoding (SSPE) tailored for HSI data, ensuring continuity without the
computational burden of convolution-based approaches. Extensive experiments on
benchmark datasets demonstrate that MemFormer achieves superior classification
accuracy, outperforming state-of-the-art methods.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:43:34 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ahmad",
"Muhammad",
""
],
[
"Mazzara",
"Manuel",
""
],
[
"Distefano",
"Salvatore",
""
],
[
"Khan",
"Adil Mehmood",
""
]
] |
2504.13243 | Juan Pablo Garces Varas | Aqeel Ahmed, Juan P. Garc\'es, Manfred Lindner | Radiative Symmetry Breaking with a Scale Invariant Seesaw | 21 pages, 8 captioned figures and 5 tables | null | null | null | hep-ph | http://creativecommons.org/licenses/by/4.0/ | We study a scale invariant inverse seesaw model that radiatively generates
the electroweak scale, the Standard Model (SM) neutrino masses, and stabilizes
the electroweak vacuum. Previous studies have noted that the SM Higgs potential
and the electroweak scale can be radiatively generated via the minimal seesaw
mechanism. This scenario, called the ``Neutrino Option'', was UV-completed by a
scale invariant framework. However, these models predict singlet neutrino and
scalar masses at $10^7-10^9$ GeV, beyond any experimental reach and leave the
electroweak vacuum meta/un-stable at high energies. In this work, we propose
modifications to this framework that lower the singlet neutrino masses to
experimentally accessible scales through an inverse seesaw mechanism, while
fully stabilizing the electroweak vacuum with an additional singlet scalar. The
possibility of generating the observed baryon asymmetry of the universe via
leptogenesis is also explored.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 17:55:37 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Ahmed",
"Aqeel",
""
],
[
"Garcés",
"Juan P.",
""
],
[
"Lindner",
"Manfred",
""
]
] |
2504.13244 | William Harvey | William Harvey, Kristan Jensen, Takahiro Uzu | Comparing top-down and bottom-up holographic defects and boundaries | 43 pages, 13 figures | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work we consider domain walls and end-of-the-world branes in AdS/CFT,
holographically dual to codimension-one conformal defects and conformal
boundaries respectively. In this setting there is an analogue of the ``bulk
point'' singularity in boundary correlation functions, which we use to compare
top-down and bottom-up constructions of these systems. For example, for a range
of parameters the D3/D5 boundary CFT cannot be imitated by a tensionful
end-of-the-world brane coupled to Einstein gravity, and in another range it can
be modeled with a negative tension brane. Along the way we compute the central
charge $b$ for the M2/M5 boundary CFT.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 18:00:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Harvey",
"William",
""
],
[
"Jensen",
"Kristan",
""
],
[
"Uzu",
"Takahiro",
""
]
] |
2504.13245 | David Trestini | David Trestini | The Schott term in the binding energy for compact binaries on circular
orbits at fourth post-Newtonian order | 21 pages | null | null | null | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | The phasing for compact binary systems on circular orbits was obtained in
arXiv:2304.11185 at fourth-and-a-half post-Newtonian (4.5PN) order thanks to
two main ingredients: the 4PN conservative energy (associated to a nonradiative
spacetime) in terms of the orbital frequency and the 4.5PN flux in terms of the
waveform frequency (i.e., the half-frequency of the $(\ell,m)=(2,2)$ mode).
When obtaining the phasing, a key physical postulate was made: the expression
of the binding energy in terms of the waveform frequency was assumed to be
identical to the expression of the conservative energy in terms of the orbital
frequency. This postulate was necessary to ensure that the frequency evolution
obtained through the flux-balance law (which involves the binding energy) was
independent of the choice of spacetime foliation. In this work, I show that the
binding energy entering the flux-balance law differs from the 4PN conservative
energy by a 4PN pseudo-Schott term, associated with radiation-reaction effects
due to gravitational tails. Unlike the usual Schott terms (at 2.5PN, 3.5PN and
4.5PN), the pseudo-Schott term is not a total derivative and is in fact
hereditary, so it does not vanish for circular orbits. Remarkably, the binding
energy thus obtained is in perfect agreement with the one obtained using the
aforementioned physical postulate, which confirms that the 4.5PN phasing
associated to the waveform frequency computed in arXiv:2304.11185 is indeed
correct. This result is extended to the other Poincar\'e invariants, and
`thermodynamic' relations between the binding energy and angular momentum are
established. Finally, the chirp and phasing associated to the orbital frequency
are presented at 4.5PN, including horizon-absorption effects.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 18:00:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Trestini",
"David",
""
]
] |
2504.13246 | Daniel Carrera | Daniel Carrera, Abigail Davenport, Jacob B. Simon, Hans Baehr, Til
Birnstiel, Cassandra Hall, David Rea, and Sebastian Stammler | Turbulence Inhibits Planetesimal Formation in Class 0/I Disks Subject to
Infall | 20 pages, 13 figures, submitted to ApJ | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | There is growing evidence that planet formation begins early, within the
$\lesssim 1$Myr Class 0/I phase, when infall dominates disk dynamics. Our goal
is to determine if Class 0/I disks reach the conditions needed to form
planetesimals ($\sim 100$km planet building blocks) by the streaming
instability (SI). We focus on a recent suggestion that early infall causes an
``inflationary'' phase in which dust grains are advected outward. We modified
the \texttt{DustPy} code to build a 1D disk that includes dust evolution,
infall, and heating and cooling sources. We ran six models and examined the
implications for the SI, taking into account recent works on how the SI
responds to external turbulence. In line with other works, we find that grains
are advected outward, which leads to ``advection-condensation-drift'' loop that
greatly enhances the dust density at the water snowline. However, we do not see
this process at the silicate line. Instead, we find a new pile up at the edge
of the expanding disk. However, despite these localized enhancements, even a
modest amount of turbulence ($\alpha = 10^{-3}$) leaves planetesimal formation
far out of reach. The midplane dust-to-gas ratio is at least an order of
magnitude below the SI threshold, even taking into account recent results on
how dust coagulation boosts the SI. For planetesimals to form in the Class 0/I
phase may require a way to transport angular momentum without turbulence (e.g.,
disk winds) or a non-SI mechanism to form planetesimals.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 18:00:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Carrera",
"Daniel",
""
],
[
"Davenport",
"Abigail",
""
],
[
"Simon",
"Jacob B.",
""
],
[
"Baehr",
"Hans",
""
],
[
"Birnstiel",
"Til",
""
],
[
"Hall",
"Cassandra",
""
],
[
"Rea",
"David",
""
],
[
"Stammler",
"Sebastian",
""
]
] |
2504.13247 | Xiuyuan Zhang | Xiuyuan Zhang, Lina Necib and Denis Erkal | Darkness in the Crust: Searching for the truly "Dark" Subhalos with
Paleo-detectors | 11 pages, 5 figures. To be submitted to PRD | null | null | null | hep-ph astro-ph.CO astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Low-mass dark matter (DM) subhalos are pivotal in understanding the
small-scale structure of the universe, thereby offering a sensitive method to
discriminate between different cosmological models. In this study, we estimate
the local number density of cold DM subhalos in the solar neighborhood, and
demonstrate that their sparse distribution makes their detection via direct
detection experiments highly improbable. However, it is plausible to expect
that an $\mathcal{O}(1)$ number of subhalos could be detected by
Paleo-detectors, a proposed new technique to look for DM by reading out damage
tracks left by past DM interactions in minerals, due to their extended exposure
times. Hence, we explore how Paleo-detectors can serve as effective probes for
the properties of low-mass subhalos, $\mathcal{O}(10^{-5}-10^8) M_{\odot}$. We
find that Paleo-detectors might be able to constrain certain regions of the
subhalo mass-concentration relation (for subhalo masses of $10-10^4 M_\odot$ if
DM has a mass of $\sim5$GeV). This is a new and complementary type of study
that seeks to combine information from the particle nature of DM to that of
small scale structures.
| [
{
"version": "v1",
"created": "Thu, 17 Apr 2025 18:00:00 GMT"
}
] | 2025-04-21T00:00:00 | [
[
"Zhang",
"Xiuyuan",
""
],
[
"Necib",
"Lina",
""
],
[
"Erkal",
"Denis",
""
]
] |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.