id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.08927 | David Acuna | David Acuna, Ximing Lu, Jaehun Jung, Hyunwoo Kim, Amlan Kar, Sanja Fidler, Yejin Choi | Socratic-MCTS: Test-Time Visual Reasoning by Asking the Right Questions | null | null | null | null | cs.CV cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent research in vision-language models (VLMs) has centered around the possibility of equipping them with implicit long-form chain-of-thought reasoning -- akin to the success observed in language models -- via distillation and reinforcement learning. But what about the non-reasoning models already trained and deployed across the internet? Should we simply abandon them, or is there hope for a search mechanism that can elicit hidden knowledge and induce long reasoning traces -- without any additional training or supervision? In this paper, we explore this possibility using a Monte Carlo Tree Search (MCTS)-inspired algorithm, which injects subquestion-subanswer pairs into the model's output stream. We show that framing reasoning as a search process -- where subquestions act as latent decisions within a broader inference trajectory -- helps the model "connect the dots" between fragmented knowledge and produce extended reasoning traces in non-reasoning models. We evaluate our method across three benchmarks and observe consistent improvements. Notably, our approach yields a 2% overall improvement on MMMU-PRO, including a significant 9% gain in Liberal Arts. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:51:16 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Acuna",
"David",
""
],
[
"Lu",
"Ximing",
""
],
[
"Jung",
"Jaehun",
""
],
[
"Kim",
"Hyunwoo",
""
],
[
"Kar",
"Amlan",
""
],
[
"Fidler",
"Sanja",
""
],
[
"Choi",
"Yejin",
""
]
] |
2506.08928 | Zhongyuan Liang | Zhongyuan Liang, Zachary T. Rewolinski, Abhineet Agarwal, Tiffany M. Tang, Bin Yu | Local MDI+: Local Feature Importances for Tree-Based Models | null | null | null | null | cs.LG stat.ME stat.ML | http://creativecommons.org/licenses/by/4.0/ | Tree-based ensembles such as random forests remain the go-to for tabular data over deep learning models due to their prediction performance and computational efficiency. These advantages have led to their widespread deployment in high-stakes domains, where interpretability is essential for ensuring trustworthy predictions. This has motivated the development of popular local (i.e. sample-specific) feature importance (LFI) methods such as LIME and TreeSHAP. However, these approaches rely on approximations that ignore the model's internal structure and instead depend on potentially unstable perturbations. These issues are addressed in the global setting by MDI+, a feature importance method which exploits an equivalence between decision trees and linear models on a transformed node basis. However, the global MDI+ scores are not able to explain predictions when faced with heterogeneous individual characteristics. To address this gap, we propose Local MDI+ (LMDI+), a novel extension of the MDI+ framework to the sample specific setting. LMDI+ outperforms existing baselines LIME and TreeSHAP in identifying instance-specific signal features, averaging a 10% improvement in downstream task performance across twelve real-world benchmark datasets. It further demonstrates greater stability by consistently producing similar instance-level feature importance rankings across multiple random forest fits. Finally, LMDI+ enables local interpretability use cases, including the identification of closer counterfactuals and the discovery of homogeneous subgroups. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:51:27 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Liang",
"Zhongyuan",
""
],
[
"Rewolinski",
"Zachary T.",
""
],
[
"Agarwal",
"Abhineet",
""
],
[
"Tang",
"Tiffany M.",
""
],
[
"Yu",
"Bin",
""
]
] |
2506.08929 | Nan Liang | Zhen Huang, Xin Luo, Bin Zhang, Jianchao Feng, Puxun Wu, Yu Liu and Nan Liang | Gamma-Ray Bursts Calibrated by Using Artificial Neural Networks from the Pantheon+ Sample | 11 pages, 7 figures, 3 tables | null | null | null | astro-ph.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we calibrate the luminosity relation of gamma-ray bursts (GRBs) by Artificial Neural Networks (ANN) which is employed to analyze the Pantheon+ sample of type Ia supernovae (SNe Ia) in a manner independent of cosmological assumptions. The A219 GRB dataset are used to calibrate the Amati relation (\(E_{\rm p}\)-\(E_{\rm iso}\)) at low redshift with the ANN framework, facilitating the construction of the Hubble diagram at higher redshifts. Cosmological models are constrained with GRBs at high-redshift and the latest observational Hubble data (OHD) via a Markov Chain Monte Carlo numerical approach. For the Chevallier-Polarski-Linder (CPL) model within a flat universe, we obtain \(\Omega_{\rm m} = 0.321^{+0.078}_{-0.069}\), \(h = 0.654^{+0.053}_{-0.071}\), \(w_0 = -1.02^{+0.67}_{-0.50}\), and \(w_a = -0.98^{+0.58}_{-0.58}\) at the 1-\(\sigma\) confidence level, which indicating a preference for dark energy with potential redshift evolution (\(w_a \neq 0\)). These findings by using ANN align closely with those derived from GRBs calibrated by using Gaussian Processes. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:52:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Huang",
"Zhen",
""
],
[
"Luo",
"Xin",
""
],
[
"Zhang",
"Bin",
""
],
[
"Feng",
"Jianchao",
""
],
[
"Wu",
"Puxun",
""
],
[
"Liu",
"Yu",
""
],
[
"Liang",
"Nan",
""
]
] |
2506.08930 | Hossein Rezaei | Hossein Rezaei, Mehmet Emin Arslan, George Yammine, Niels Neumann and Norman Franchi | Hardware Limitations and Optimization Approach in 1-Bit RIS Design at 28 GHz | This work has been submitted to the IEEE for possible publication | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Reconfigurable intelligent surfaces (RIS) have emerged as a transformative technology for electromagnetic (EM) wave manipulation, offering unprecedented control over wave reflections compared to traditional metallic reflectors. By utilizing an array of tunable elements, RIS can steer and shape electromagnetic waves to enhance signal quality in wireless communication and radar systems. However, practical implementations face significant challenges due to hardware limitations and phase quantization errors. In this work, a 1-bit RIS prototype operating at 28 GHz is developed to experimentally evaluate the impact of hardware constraints on RIS performance. Unlike conventional studies that model RIS as an ideal phase-shift matrix, this study accounts for physical parameters that influence the actual reflection pattern. In particular, the presence of specular reflection due to hardware limitations is investigated. Additionally, the effects of phase quantization errors, which stem from the discrete nature of RIS elements, are analyzed, and a genetic algorithm (GA)-based optimization is introduced to mitigate these errors. The proposed optimization strategy effectively reduces gain degradation at the desired angle caused by 1-bit quantization, enhancing the overall performance of RIS. The effectiveness of the approach is validated through measurements, underscoring the importance of advanced phase control techniques in improving the functionality of RIS. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:57:17 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Rezaei",
"Hossein",
""
],
[
"Arslan",
"Mehmet Emin",
""
],
[
"Yammine",
"George",
""
],
[
"Neumann",
"Niels",
""
],
[
"Franchi",
"Norman",
""
]
] |
2506.08931 | Yixuan Li | Yixuan Li, Yutang Lin, Jieming Cui, Tengyu Liu, Wei Liang, Yixin Zhu, Siyuan Huang | CLONE: Closed-Loop Whole-Body Humanoid Teleoperation for Long-Horizon Tasks | 18 pages, 13 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Humanoid teleoperation plays a vital role in demonstrating and collecting data for complex humanoid-scene interactions. However, current teleoperation systems face critical limitations: they decouple upper- and lower-body control to maintain stability, restricting natural coordination, and operate open-loop without real-time position feedback, leading to accumulated drift. The fundamental challenge is achieving precise, coordinated whole-body teleoperation over extended durations while maintaining accurate global positioning. Here we show that an MoE-based teleoperation system, CLONE, with closed-loop error correction enables unprecedented whole-body teleoperation fidelity, maintaining minimal positional drift over long-range trajectories using only head and hand tracking from an MR headset. Unlike previous methods that either sacrifice coordination for stability or suffer from unbounded drift, CLONE learns diverse motion skills while preventing tracking error accumulation through real-time feedback, enabling complex coordinated movements such as ``picking up objects from the ground.'' These results establish a new milestone for whole-body humanoid teleoperation for long-horizon humanoid-scene interaction tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:58:29 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Li",
"Yixuan",
""
],
[
"Lin",
"Yutang",
""
],
[
"Cui",
"Jieming",
""
],
[
"Liu",
"Tengyu",
""
],
[
"Liang",
"Wei",
""
],
[
"Zhu",
"Yixin",
""
],
[
"Huang",
"Siyuan",
""
]
] |
2506.08932 | Haochen Wang | Haochen Wang, Kiyoshi Masui, Shion Andrew, Emmanuel Fonseca, B. M. Gaensler, R. C. Joseph, Victoria M. Kaspi, Bikash Kharel, Adam E. Lanman, Calvin Leung, Lluis Mas-Ribas, Juan Mena-Parra, Kenzie Nimmo, Aaron B. Pearlman, Ue-Li Pen, J. Xavier Prochaska, Ryan Raikman, Kaitlyn Shin, Seth R. Siegel, Kendrick M. Smith, Ingrid H. Stairs | Measurement of the Dispersion$\unicode{x2013}$Galaxy Cross-Power Spectrum with the Second CHIME/FRB Catalog | null | null | null | null | astro-ph.CO astro-ph.GA astro-ph.HE astro-ph.IM gr-qc | http://creativecommons.org/licenses/by/4.0/ | The dispersion of extragalactic fast radio bursts (FRBs) can serve as a powerful probe of the diffuse plasma between and surrounding galaxies, which contains most of the Universe's baryons. By cross-correlating the dispersion of background FRBs with the locations of foreground galaxies, we can study the relative spatial distributions of plasma and galaxies on scales of 0.1 to 50 Mpc, which are strongly affected by feedback processes in galaxy formation. Here we present the measurement of the dispersion$\unicode{x2013}$galaxy angular cross-power spectrum between 2873 FRBs from the Second CHIME/FRB Catalog and nearly 6 million galaxies from the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Survey. Over five photometric galaxy redshift bins spanning $0.05 < z <0.5$ and at 5.1$\sigma$ significance, we make the first definitive detection of spatial correlations in FRB dispersion measure due to cosmic structure. While parameter inferences should be interpreted with caution because of incomplete modelling of both the signal and systematic errors, our data indicate that the plasma$\unicode{x2013}$galaxy cross-power spectrum cuts off relative to the matter power spectrum at a scale $k_\textrm{cut}^{-1}=0.9^{+0.4}_{-0.4}\,\textrm{Mpc}$. This scale is consistent with those X-ray stacking analyses that suggest dark-matter halos with group-scale masses are largely evacuated of their baryons by feedback processes. Our study demonstrates that FRBs are promising tools to discern the physics of baryonic structure formation and will only become more powerful as FRB surveys expand. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:59:28 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Haochen",
""
],
[
"Masui",
"Kiyoshi",
""
],
[
"Andrew",
"Shion",
""
],
[
"Fonseca",
"Emmanuel",
""
],
[
"Gaensler",
"B. M.",
""
],
[
"Joseph",
"R. C.",
""
],
[
"Kaspi",
"Victoria M.",
""
],
[
"Kharel",
"Bikash",
""
],
[
"Lanman",
"Adam E.",
""
],
[
"Leung",
"Calvin",
""
],
[
"Mas-Ribas",
"Lluis",
""
],
[
"Mena-Parra",
"Juan",
""
],
[
"Nimmo",
"Kenzie",
""
],
[
"Pearlman",
"Aaron B.",
""
],
[
"Pen",
"Ue-Li",
""
],
[
"Prochaska",
"J. Xavier",
""
],
[
"Raikman",
"Ryan",
""
],
[
"Shin",
"Kaitlyn",
""
],
[
"Siegel",
"Seth R.",
""
],
[
"Smith",
"Kendrick M.",
""
],
[
"Stairs",
"Ingrid H.",
""
]
] |
2506.08933 | Wendong Bu | Wendong Bu, Yang Wu, Qifan Yu, Minghe Gao, Bingchen Miao, Zhenkui Zhang, Kaihang Pan, Yunfei Li, Mengze Li, Wei Ji, Juncheng Li, Siliang Tang, Yueting Zhuang | What Limits Virtual Agent Application? OmniBench: A Scalable Multi-Dimensional Benchmark for Essential Virtual Agent Capabilities | Accepted by ICML 2025 (Oral) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As multimodal large language models (MLLMs) advance, MLLM-based virtual agents have demonstrated remarkable performance. However, existing benchmarks face significant limitations, including uncontrollable task complexity, extensive manual annotation with limited scenarios, and a lack of multidimensional evaluation. In response to these challenges, we introduce OmniBench, a self-generating, cross-platform, graph-based benchmark with an automated pipeline for synthesizing tasks of controllable complexity through subtask composition. To evaluate the diverse capabilities of virtual agents on the graph, we further present OmniEval, a multidimensional evaluation framework that includes subtask-level evaluation, graph-based metrics, and comprehensive tests across 10 capabilities. Our synthesized dataset contains 36k graph-structured tasks across 20 scenarios, achieving a 91\% human acceptance rate. Training on our graph-structured data shows that it can more efficiently guide agents compared to manually annotated data. We conduct multidimensional evaluations for various open-source and closed-source models, revealing their performance across various capabilities and paving the way for future advancements. Our project is available at https://omni-bench.github.io/. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 15:59:38 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bu",
"Wendong",
""
],
[
"Wu",
"Yang",
""
],
[
"Yu",
"Qifan",
""
],
[
"Gao",
"Minghe",
""
],
[
"Miao",
"Bingchen",
""
],
[
"Zhang",
"Zhenkui",
""
],
[
"Pan",
"Kaihang",
""
],
[
"Li",
"Yunfei",
""
],
[
"Li",
"Mengze",
""
],
[
"Ji",
"Wei",
""
],
[
"Li",
"Juncheng",
""
],
[
"Tang",
"Siliang",
""
],
[
"Zhuang",
"Yueting",
""
]
] |
2506.08935 | Andrew Shin | Andrew Shin | Can A Gamer Train A Mathematical Reasoning Model? | null | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | While large language models (LLMs) have achieved remarkable performance in various tasks including mathematical reasoning, their development typically demands prohibitive computational resources. Recent advancements have reduced costs for training capable models, yet even these approaches rely on high-end hardware clusters. In this paper, we demonstrate that a single average gaming GPU can train a solid mathematical reasoning model, by integrating reinforcement learning and memory optimization techniques. Specifically, we train a 1.5B parameter mathematical reasoning model on RTX 3080 Ti of 16GB memory that achieves comparable or better performance on mathematical reasoning benchmarks than models several times larger, in resource-constrained environments. Our results challenge the paradigm that state-of-the-art mathematical reasoning necessitates massive infrastructure, democratizing access to high-performance AI research. https://github.com/shinandrew/YouronMath. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:00:12 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Shin",
"Andrew",
""
]
] |
2506.08936 | Amina Mollaysa | Amina Mollaysa, Artem Moskale, Pushpak Pati, Tommaso Mansi, Mangal Prakash, Rui Liao | BioLangFusion: Multimodal Fusion of DNA, mRNA, and Protein Language Models | Proceedings of ICML 2025 Workshop on Multi-modal Foundation Proceedings of ICML 2025 Workshop on Multi-modal Foundation Proceedings of ICML 2025 Workshop on Multi-modal Foundation Models and Large Language Models for Life Sciences | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We present BioLangFusion, a simple approach for integrating pre-trained DNA, mRNA, and protein language models into unified molecular representations. Motivated by the central dogma of molecular biology (information flow from gene to transcript to protein), we align per-modality embeddings at the biologically meaningful codon level (three nucleotides encoding one amino acid) to ensure direct cross-modal correspondence. BioLangFusion studies three standard fusion techniques: (i) codon-level embedding concatenation, (ii) entropy-regularized attention pooling inspired by multiple-instance learning, and (iii) cross-modal multi-head attention -- each technique providing a different inductive bias for combining modality-specific signals. These methods require no additional pre-training or modification of the base models, allowing straightforward integration with existing sequence-based foundation models. Across five molecular property prediction tasks, BioLangFusion outperforms strong unimodal baselines, showing that even simple fusion of pre-trained models can capture complementary multi-omic information with minimal overhead. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:01:11 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Mollaysa",
"Amina",
""
],
[
"Moskale",
"Artem",
""
],
[
"Pati",
"Pushpak",
""
],
[
"Mansi",
"Tommaso",
""
],
[
"Prakash",
"Mangal",
""
],
[
"Liao",
"Rui",
""
]
] |
2506.08937 | Diancong Jin | Diancong Jin | Asymptotic error distribution for stochastic Runge--Kutta methods of strong order one | null | null | null | null | math.NA cs.NA math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work gives the asymptotic error distribution of the stochastic Runge--Kutta (SRK) method of strong order $1$ applied to Stratonovich-type stochastic differential equations. For dealing with the implicitness introduced in the diffusion term, we provide a framework to derive the asymptotic error distribution of diffusion-implicit or fully implicit numerical methods, which enables us to construct a fully explicit numerical method sharing the same asymptotic error distribution as the SRK method. Further, we show for the multiplicative noise that the limit distribution $U(T)$ satisfies $\mathbf E|U(T)|^2\le e^{L_1T}(1+\eta_1)T^3$ for some $\eta_1$ only depending on the coefficients of the SRK method. Thus, we infer that $\eta_1$ is the key parameter reflecting the growth rate of the mean-square error of the SRK method. Especially, among the SRK methods of strong order $1$, those of weak order $2$ ($\eta_1=0$) share the unified asymptotic error distribution and have the smallest mean-square errors after a long time. This property is also found for the case of additive noise. It seems that we are the first to give the asymptotic error distribution of fully implicit numerical methods for SDEs. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:01:50 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Jin",
"Diancong",
""
]
] |
2506.08938 | Qinggang Zhang | Qinggang Zhang, Zhishang Xiang, Yilin Xiao, Le Wang, Junhui Li, Xinrun Wang, and Jinsong Su | FaithfulRAG: Fact-Level Conflict Modeling for Context-Faithful Retrieval-Augmented Generation | Qinggang Zhang and Zhishang Xiang contributed equally to this work. Corresponding author: Jinsong Su | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) augmented with retrieval systems have demonstrated significant potential in handling knowledge-intensive tasks. However, these models often struggle with unfaithfulness issues, generating outputs that either ignore the retrieved context or inconsistently blend it with the LLM`s parametric knowledge. This issue is particularly severe in cases of knowledge conflict, where the retrieved context conflicts with the model`s parametric knowledge. While existing faithful RAG approaches enforce strict context adherence through well-designed prompts or modified decoding strategies, our analysis reveals a critical limitation: they achieve faithfulness by forcibly suppressing the model`s parametric knowledge, which undermines the model`s internal knowledge structure and increases the risk of misinterpreting the context. To this end, this paper proposes FaithfulRAG, a novel framework that resolves knowledge conflicts by explicitly modeling discrepancies between the model`s parametric knowledge and retrieved context. Specifically, FaithfulRAG identifies conflicting knowledge at the fact level and designs a self-thinking process, allowing LLMs to reason about and integrate conflicting facts before generating responses. Extensive experiments demonstrate that our method outperforms state-of-the-art methods. The code is available at https:// github.com/DeepLearnXMU/Faithful-RAG | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:02:54 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhang",
"Qinggang",
""
],
[
"Xiang",
"Zhishang",
""
],
[
"Xiao",
"Yilin",
""
],
[
"Wang",
"Le",
""
],
[
"Li",
"Junhui",
""
],
[
"Wang",
"Xinrun",
""
],
[
"Su",
"Jinsong",
""
]
] |
2506.08939 | Ye Hang | Hang Ye, Gaoxiang Duan, Haoran Zeng, Yangxin Zhu, Lingxue Meng, Xiaoying Zheng and Yongxin Zhu | KARMA: A Multilevel Decomposition Hybrid Mamba Framework for Multivariate Long-Term Time Series Forecasting | 10 pages,3 figures, published to WASA2025 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multivariate long-term and efficient time series forecasting is a key requirement for a variety of practical applications, and there are complex interleaving time dynamics in time series data that require decomposition modeling. Traditional time series decomposition methods are single and rely on fixed rules, which are insufficient for mining the potential information of the series and adapting to the dynamic characteristics of complex series. On the other hand, the Transformer-based models for time series forecasting struggle to effectively model long sequences and intricate dynamic relationships due to their high computational complexity. To overcome these limitations, we introduce KARMA, with an Adaptive Time Channel Decomposition module (ATCD) to dynamically extract trend and seasonal components. It further integrates a Hybrid Frequency-Time Decomposition module (HFTD) to further decompose Series into frequency-domain and time-domain. These components are coupled with multi-scale Mamba-based KarmaBlock to efficiently process global and local information in a coordinated manner. Experiments on eight real-world datasets from diverse domains well demonstrated that KARMA significantly outperforms mainstream baseline methods in both predictive accuracy and computational efficiency. Code and full results are available at this repository: https://github.com/yedadasd/KARMA | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:03:33 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Ye",
"Hang",
""
],
[
"Duan",
"Gaoxiang",
""
],
[
"Zeng",
"Haoran",
""
],
[
"Zhu",
"Yangxin",
""
],
[
"Meng",
"Lingxue",
""
],
[
"Zheng",
"Xiaoying",
""
],
[
"Zhu",
"Yongxin",
""
]
] |
2506.08940 | Adam Bednorz | Adam Bednorz, Josep Batle, Tomasz Bia{\l}ecki, Jaros{\l}aw K. Korbicz | Closing objectivity loophole in Bell tests on a public quantum computer | 13 pages, 12 figures, 7 tables | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We have constructed and run a Bell test of local realism focusing on the objectivity criterion. The objectivity means that the outcomes are confirmed macroscopically by a few observers at each of party. The IBM Quantum devices turn out to be sufficiently accurate to pass such an extended Bell-type test, although at the price of communication loopholes and residual but statistically significant signaling. The test serves also as the benchmark of entanglement spread across larger sets of qubits. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:03:39 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bednorz",
"Adam",
""
],
[
"Batle",
"Josep",
""
],
[
"Białecki",
"Tomasz",
""
],
[
"Korbicz",
"Jarosław K.",
""
]
] |
2506.08941 | Piotr Rozmej | Piotr Rozmej and Anna Karczewska | Solitary wave solutions, periodic and superposition solutions to the system of first-order (2+1)-dimensional Boussinesq's equations derived from the Euler equations for an ideal fluid model | 32 pages, 8 Figures | null | null | null | nlin.PS math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article concludes the study of (2+1)-dimensional nonlinear wave equations that can be derived in a model of an ideal fluid with irrotational motion. In the considered case of identical scaling of the $x,y$ variables, obtaining a (2+1)-dimensional wave equation analogous to the KdV equation is impossible. Instead, from a system of two first-order Boussinesq equations, a non-linear wave equation for the auxiliary function $f(x,y,z)$ defining the velocity potential can be obtained, and only from its solutions can the surface wave form $\eta(x,y,z)$ be obtained. We demonstrate the existence of families of (2+1)-dimensional traveling wave solutions, including solitary and periodic solutions, of both cnoidal and superposition types. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:04:55 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Rozmej",
"Piotr",
""
],
[
"Karczewska",
"Anna",
""
]
] |
2506.08943 | Mohammadrasoul Taghavi | Edwin A. Marengo, Mohammadrasoul Taghavi | Equivalence of Optical Theorems | 18 pages, 6 figures | null | null | null | physics.optics math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We demonstrate, in the full vector formulation of electromagnetic fields, that the well-known optical theorem pertinent for the characterization of a scatterer's extinction power and associated cross section can be expressed in a multitude of alternative equivalent forms. These alternatives involve different forms of projective field measurements or detectors. The inherent nonuniqueness of such optical-theorem-based detectors stems from the nonuniqueness of an associated inverse source problem, and can be interpreted via well-known equivalence principles. Some of the multiple ways in which the extinction of power due to the interaction of a scattering body with a probing field can be measured remotely are derived and interpreted for a number of canonical frameworks. This includes detectors and their corresponding optical theorems synthesized in the contexts of surface-confined sensors for near-field sensing, surface sensors based on backpropagation-based imaging, a number of planar aperture realizations dealt with through classical diffraction theory, as well as detectors based on multipole representations. General aspects of the derived optical theorems are discussed in the context of envisioned practical applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:05:37 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Marengo",
"Edwin A.",
""
],
[
"Taghavi",
"Mohammadrasoul",
""
]
] |
2506.08944 | Ahmad Halimi Razlighi | Ahmad Halimi Razlighi, Carsten Bockelmann, and Armin Dekorsy | Semantic Communication for Cooperative Multi-Tasking over Rate-Limited Wireless Channels with Implicit Optimal Prior | This work has been submitted to the IEEE for possible publication | null | null | null | eess.SP cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we expand the cooperative multi-task semantic communication framework (CMT-SemCom) introduced in [1], which divides the semantic encoder on the transmitter side into a common unit (CU) and multiple specific units (SUs), to a more applicable design. Our proposed system model addresses real-world constraints by introducing a general design that operates over rate-limited wireless channels. Further, we aim to tackle the rate-limit constraint, represented through the Kullback-Leibler (KL) divergence, by employing the density ratio trick alongside the implicit optimal prior method (IoPm). By applying the IoPm to our multi-task processing framework, we propose a hybrid learning approach that combines deep neural networks with kernelized-parametric machine learning methods, enabling a robust solution for the CMT-SemCom. Our framework is grounded in information-theoretic principles and employs variational approximations to bridge theoretical foundations with practical implementations. Simulation results demonstrate the proposed system's effectiveness in rate-constrained multi-task SemCom scenarios, highlighting its potential for enabling intelligence in next-generation wireless networks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:05:54 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Razlighi",
"Ahmad Halimi",
""
],
[
"Bockelmann",
"Carsten",
""
],
[
"Dekorsy",
"Armin",
""
]
] |
2506.08945 | Simone Daniotti | Simone Daniotti, Johannes Wachs, Xiangnan Feng, Frank Neffke | Who is using AI to code? Global diffusion and impact of generative AI | 42 pages | null | null | null | cs.CY cs.SE physics.soc-ph | http://creativecommons.org/licenses/by/4.0/ | Generative coding tools promise big productivity gains, but uneven uptake could widen skill and income gaps. We train a neural classifier to spot AI-generated Python functions in 80 million GitHub commits (2018-2024) by 200,000 developers and track how fast--and where--these tools take hold. By December 2024, AI wrote an estimated 30.1% of Python functions from U.S. contributors, versus 24.3% in Germany, 23.2% in France, 21.6% in India, 15.4% in Russia and 11.7% in China. Newer GitHub users use AI more than veterans, while male and female developers adopt at similar rates. Within-developer fixed-effects models show that moving to 30% AI use raises quarterly commits by 2.4%. Coupling this effect with occupational task and wage data puts the annual value of AI-assisted coding in the United States at $9.6-$14.4 billion, rising to $64-$96 billion if we assume higher estimates of productivity effects reported by randomized control trials. Moreover, generative AI prompts learning and innovation, leading to increases in the number of new libraries and library combinations that programmers use. In short, AI usage is already widespread but highly uneven, and the intensity of use, not only access, drives measurable gains in output and exploration. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:06:19 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Daniotti",
"Simone",
""
],
[
"Wachs",
"Johannes",
""
],
[
"Feng",
"Xiangnan",
""
],
[
"Neffke",
"Frank",
""
]
] |
2506.08946 | Jesus Javier Cobos-Martinez | J. J. Cobos-Mart\'inez, Guilherme N. Zeminiani, Kazuo Tsushima | Heavy-heavy and heavy-light mesons in cold nuclear matter | Published, Review article, 56 pages, 20 figures | Symmetry 2025, 17(5), 787 | 10.3390/sym17050787 | null | nucl-th hep-ex hep-ph nucl-ex | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We review the in-medium modifications of effective masses (Lorentz scalar potentials or phenomenon of mass shift) of heavy-heavy and heavy-light mesons in symmetric nuclear matter and their nuclear bound states. We use a combined approach with the quark-meson coupling (QMC) model and an effective Lagrangian. As demonstrated by the cases of pionic and kaonic atoms, studies of meson-nucleus bound state can provide us with important information on chiral symmetry in dense nuclear medium. In this review, we treat the mesons, $K, K^*, D, D^*, B, B^*, \eta, \eta', \phi, \eta_c, J/\psi, \eta_b, \Upsilon$, and $B_c$, where our emphasys is on the heavy mesons. In addition, we also present some new results for the $B_c$-nucleus bound states. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:08:24 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Cobos-Martínez",
"J. J.",
""
],
[
"Zeminiani",
"Guilherme N.",
""
],
[
"Tsushima",
"Kazuo",
""
]
] |
2506.08947 | Vicente P. Soloviev | Vicente P. Soloviev and Antonio M\'arquez Romero and Josh Kirsopp and Michal Krompiec | Scaling Portfolio Diversification with Quantum Circuit Cutting Techniques | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quantum Approximate Optimization Algorithms (QAOA) have demonstrated a strong potential in addressing graph-based optimization problems. However, the execution of large-scale quantum circuits remains constrained by the limitations of current quantum hardware. In this work, we introduce QuantCut, an automatic framework for circuit cutting that enables efficient execution of large quantum circuits by decomposing entangling two-qubit gates into manageable sub-circuits. Specifically, we focus on gate-cutting techniques. We apply QuantCut to a 71-qubit QAOA circuit ansatz for portfolio diversification in the S&P 500 stock market, aiming to maximize asset diversification. Our approach iteratively optimizes the expectation value while leveraging circuit-cutting strategies to reduce the qubit register size. To validate our framework, we first conduct experiments on a toy model using quantum noise simulations for the Max-Cut problem, analyzing performance improvements with an increasing number of layers. Subsequently, we extend our methodology to a real-world financial optimization scenario, showing competitive results. The results suggest that QuantCut effectively facilitates large-scale quantum computations with circuit-cutting technologies. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:08:29 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Soloviev",
"Vicente P.",
""
],
[
"Romero",
"Antonio Márquez",
""
],
[
"Kirsopp",
"Josh",
""
],
[
"Krompiec",
"Michal",
""
]
] |
2506.08948 | Alessandro Fregosi | Alessandro Fregosi, Fernando Brandi, Luca Labate, Federica Baffigi, Gianluca Cellamare, Mohamed Ezzat, Daniele Palla, Guido Toci, Alex Whitehead and Leonida A. Gizzi | A study of cross-relaxation and temporal dynamics of lasing at 2 microns in Thulium doped ceramic | null | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | We report the characterization of the pump absorption and emission dynamic properties of a \tulio{} ceramic lasing medium using a three mirrors folded laser cavity. We measured a slope efficiency of 73\%, which allowed us to retrieve the cross-relaxation coefficient. The behavior of our system was modeled via a set of macroscopic rate equations in both the quasi continuous wave and the pulsed pumping regime. Numerical solutions were obtained, showing a good agreement with the experimental findings. The numerical solution also yielded a cross-relaxation coefficient in very good agreement with the measured one, showing that the cross-relaxation phenomenon approaches the maximum theoretical efficiency. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:09:16 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Fregosi",
"Alessandro",
""
],
[
"Brandi",
"Fernando",
""
],
[
"Labate",
"Luca",
""
],
[
"Baffigi",
"Federica",
""
],
[
"Cellamare",
"Gianluca",
""
],
[
"Ezzat",
"Mohamed",
""
],
[
"Palla",
"Daniele",
""
],
[
"Toci",
"Guido",
""
],
[
"Whitehead",
"Alex",
""
],
[
"Gizzi",
"Leonida A.",
""
]
] |
2506.08949 | Zeyu Zhang | Hongjie Zhu, Xiwei Liu, Rundong Xue, Zeyu Zhang, Yong Xu, Daji Ergu, Ying Cai, Yang Zhao | SSS: Semi-Supervised SAM-2 with Efficient Prompting for Medical Imaging Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In the era of information explosion, efficiently leveraging large-scale unlabeled data while minimizing the reliance on high-quality pixel-level annotations remains a critical challenge in the field of medical imaging. Semi-supervised learning (SSL) enhances the utilization of unlabeled data by facilitating knowledge transfer, significantly improving the performance of fully supervised models and emerging as a highly promising research direction in medical image analysis. Inspired by the ability of Vision Foundation Models (e.g., SAM-2) to provide rich prior knowledge, we propose SSS (Semi-Supervised SAM-2), a novel approach that leverages SAM-2's robust feature extraction capabilities to uncover latent knowledge in unlabeled medical images, thus effectively enhancing feature support for fully supervised medical image segmentation. Specifically, building upon the single-stream "weak-to-strong" consistency regularization framework, this paper introduces a Discriminative Feature Enhancement (DFE) mechanism to further explore the feature discrepancies introduced by various data augmentation strategies across multiple views. By leveraging feature similarity and dissimilarity across multi-scale augmentation techniques, the method reconstructs and models the features, thereby effectively optimizing the salient regions. Furthermore, a prompt generator is developed that integrates Physical Constraints with a Sliding Window (PCSW) mechanism to generate input prompts for unlabeled data, fulfilling SAM-2's requirement for additional prompts. Extensive experiments demonstrate the superiority of the proposed method for semi-supervised medical image segmentation on two multi-label datasets, i.e., ACDC and BHSD. Notably, SSS achieves an average Dice score of 53.15 on BHSD, surpassing the previous state-of-the-art method by +3.65 Dice. Code will be available at https://github.com/AIGeeksGroup/SSS. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:09:40 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhu",
"Hongjie",
""
],
[
"Liu",
"Xiwei",
""
],
[
"Xue",
"Rundong",
""
],
[
"Zhang",
"Zeyu",
""
],
[
"Xu",
"Yong",
""
],
[
"Ergu",
"Daji",
""
],
[
"Cai",
"Ying",
""
],
[
"Zhao",
"Yang",
""
]
] |
2506.08950 | Mengqi Li | Mengqi Li | Fragility in Average Treatment Effect on the Treated under Limited Covariate Support | null | null | null | null | econ.EM | http://creativecommons.org/licenses/by/4.0/ | This paper studies the identification of the average treatment effect on the treated (ATT) under unconfoundedness when covariate overlap is partial. A formal diagnostic is proposed to characterize empirical support -- the subset of the covariate space where ATT is point-identified due to the presence of comparable untreated units. Where support is absent, standard estimators remain computable but cease to identify meaningful causal parameters. A general sensitivity framework is developed, indexing identified sets by curvature constraints on the selection mechanism. This yields a structural selection frontier tracing the trade-off between assumption strength and inferential precision. Two diagnostic statistics are introduced: the minimum assumption strength for sign identification (MAS-SI), and a fragility index that quantifies the minimal deviation from ignorability required to overturn qualitative conclusions. Applied to the LaLonde (1986) dataset, the framework reveals that nearly half the treated strata lack empirical support, rendering the ATT undefined in those regions. Simulations confirm that ATT estimates may be stable in magnitude yet fragile in epistemic content. These findings reframe overlap not as a regularity condition but as a prerequisite for identification, and recast sensitivity analysis as integral to empirical credibility rather than auxiliary robustness. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:13:21 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Li",
"Mengqi",
""
]
] |
2506.08951 | Sipra Hota | Sipra Hota, Richard de Grijs and Annapurni Subramaniam | UVIT Study of the Magellanic Clouds (U-SMAC). III. Hierarchical Star Formation in the Small Magellanic Cloud Regulated by Turbulence | 15 pages, 8 figures, 2 tables, Accepted for the publication in Astrophysical Journal | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The Small Magellanic Cloud (SMC), a satellite galaxy of the Milky Way, is an irregular dwarf galaxy exhibiting evidence of recent and ongoing star formation. We performed a spatial clustering analysis of far-ultraviolet stars in the SMC younger than 150 Myr using data from the Ultra Violet Imaging Telescope onboard AstroSat. We identified 236 young stellar structures as surface overdensities at different significance levels. The sizes of these structures range from a few parsecs to several hundred parsecs. Their irregular morphologies are characterized by a perimeter-area dimension, derived from the projected boundaries of the young stellar structures, of Dp = 1.46 +/- 0.4. The 2D fractal dimensions obtained from, respectively, the number-size relation and the size distribution are D2 = 1.64 +/- 0.03 and D2 = 1.31 +/- 0.16. These values indicate significant lumpiness among the young stellar structures. In addition, the surface density distribution of the identified structures follows a log-normal distribution. These features are strikingly similar to those of the turbulent interstellar medium, thus supporting the scenario of hierarchical star formation regulated by supersonic turbulence. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:17:04 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Hota",
"Sipra",
""
],
[
"de Grijs",
"Richard",
""
],
[
"Subramaniam",
"Annapurni",
""
]
] |
2506.08953 | Anirudh Nanduri | Anirudh Nanduri, Siyuan Huang, Rama Chellappa | Cross-Spectral Body Recognition with Side Information Embedding: Benchmarks on LLCM and Analyzing Range-Induced Occlusions on IJB-MDF | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Vision Transformers (ViTs) have demonstrated impressive performance across a wide range of biometric tasks, including face and body recognition. In this work, we adapt a ViT model pretrained on visible (VIS) imagery to the challenging problem of cross-spectral body recognition, which involves matching images captured in the visible and infrared (IR) domains. Recent ViT architectures have explored incorporating additional embeddings beyond traditional positional embeddings. Building on this idea, we integrate Side Information Embedding (SIE) and examine the impact of encoding domain and camera information to enhance cross-spectral matching. Surprisingly, our results show that encoding only camera information - without explicitly incorporating domain information - achieves state-of-the-art performance on the LLCM dataset. While occlusion handling has been extensively studied in visible-spectrum person re-identification (Re-ID), occlusions in visible-infrared (VI) Re-ID remain largely underexplored - primarily because existing VI-ReID datasets, such as LLCM, SYSU-MM01, and RegDB, predominantly feature full-body, unoccluded images. To address this gap, we analyze the impact of range-induced occlusions using the IARPA Janus Benchmark Multi-Domain Face (IJB-MDF) dataset, which provides a diverse set of visible and infrared images captured at various distances, enabling cross-range, cross-spectral evaluations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:20:52 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Nanduri",
"Anirudh",
""
],
[
"Huang",
"Siyuan",
""
],
[
"Chellappa",
"Rama",
""
]
] |
2506.08954 | Pascal Notin | Ruben Weitzman, Peter M{\o}rch Groth, Lood Van Niekerk, Aoi Otani, Yarin Gal, Debora Marks, Pascal Notin | Protriever: End-to-End Differentiable Protein Homology Search for Fitness Prediction | Accepted at ICML 2025 | null | null | null | q-bio.QM cs.LG q-bio.BM | http://creativecommons.org/licenses/by/4.0/ | Retrieving homologous protein sequences is essential for a broad range of protein modeling tasks such as fitness prediction, protein design, structure modeling, and protein-protein interactions. Traditional workflows have relied on a two-step process: first retrieving homologs via Multiple Sequence Alignments (MSA), then training models on one or more of these alignments. However, MSA-based retrieval is computationally expensive, struggles with highly divergent sequences or complex insertions & deletions patterns, and operates independently of the downstream modeling objective. We introduce Protriever, an end-to-end differentiable framework that learns to retrieve relevant homologs while simultaneously training for the target task. When applied to protein fitness prediction, Protriever achieves state-of-the-art performance compared to sequence-based models that rely on MSA-based homolog retrieval, while being two orders of magnitude faster through efficient vector search. Protriever is both architecture- and task-agnostic, and can flexibly adapt to different retrieval strategies and protein databases at inference time -- offering a scalable alternative to alignment-centric approaches. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:24:09 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Weitzman",
"Ruben",
""
],
[
"Groth",
"Peter Mørch",
""
],
[
"Van Niekerk",
"Lood",
""
],
[
"Otani",
"Aoi",
""
],
[
"Gal",
"Yarin",
""
],
[
"Marks",
"Debora",
""
],
[
"Notin",
"Pascal",
""
]
] |
2506.08955 | Chunming He | Chunming He, Kai Li, Yachao Zhang, Ziyun Yang, Youwei Pang, Longxiang Tang, Chengyu Fang, Yulun Zhang, Linghe Kong, Xiu Li, Sina Farsiu | Segment Concealed Objects with Incomplete Supervision | IEEE TPAMI | 10.1109/TPAMI.2025.3576209 | null | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Incompletely-Supervised Concealed Object Segmentation (ISCOS) involves segmenting objects that seamlessly blend into their surrounding environments, utilizing incompletely annotated data, such as weak and semi-annotations, for model training. This task remains highly challenging due to (1) the limited supervision provided by the incompletely annotated training data, and (2) the difficulty of distinguishing concealed objects from the background, which arises from the intrinsic similarities in concealed scenarios. In this paper, we introduce the first unified method for ISCOS to address these challenges. To tackle the issue of incomplete supervision, we propose a unified mean-teacher framework, SEE, that leverages the vision foundation model, ``\emph{Segment Anything Model (SAM)}'', to generate pseudo-labels using coarse masks produced by the teacher model as prompts. To mitigate the effect of low-quality segmentation masks, we introduce a series of strategies for pseudo-label generation, storage, and supervision. These strategies aim to produce informative pseudo-labels, store the best pseudo-labels generated, and select the most reliable components to guide the student model, thereby ensuring robust network training. Additionally, to tackle the issue of intrinsic similarity, we design a hybrid-granularity feature grouping module that groups features at different granularities and aggregates these results. By clustering similar features, this module promotes segmentation coherence, facilitating more complete segmentation for both single-object and multiple-object images. We validate the effectiveness of our approach across multiple ISCOS tasks, and experimental results demonstrate that our method achieves state-of-the-art performance. Furthermore, SEE can serve as a plug-and-play solution, enhancing the performance of existing models. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:25:15 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"He",
"Chunming",
""
],
[
"Li",
"Kai",
""
],
[
"Zhang",
"Yachao",
""
],
[
"Yang",
"Ziyun",
""
],
[
"Pang",
"Youwei",
""
],
[
"Tang",
"Longxiang",
""
],
[
"Fang",
"Chengyu",
""
],
[
"Zhang",
"Yulun",
""
],
[
"Kong",
"Linghe",
""
],
[
"Li",
"Xiu",
""
],
[
"Farsiu",
"Sina",
""
]
] |
2506.08956 | DaeEun Yoon | DaeEun Yoon, Semin Kim, SangWook Yoo, Jongha Lee | Data Augmentation For Small Object using Fast AutoAugment | Accepted and published in the USB Proceedings of the 20th International Conference on Modeling Decisions for Artificial Intelligence (MDAI 2023), Ume{\aa}, Sweden, June 19--22, 2023, ISBN 978-91-527-7293-5, pp.\ 12--21 | null | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, there has been tremendous progress in object detection performance. However, despite these advances, the detection performance for small objects is significantly inferior to that of large objects. Detecting small objects is one of the most challenging and important problems in computer vision. To improve the detection performance for small objects, we propose an optimal data augmentation method using Fast AutoAugment. Through our proposed method, we can quickly find optimal augmentation policies that can overcome degradation when detecting small objects, and we achieve a 20% performance improvement on the DOTA dataset. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:25:40 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Yoon",
"DaeEun",
""
],
[
"Kim",
"Semin",
""
],
[
"Yoo",
"SangWook",
""
],
[
"Lee",
"Jongha",
""
]
] |
2506.08957 | Yash Ranjan | Yash Ranjan, Rahul Sengupta, Anand Rangarajan, Sanjay Ranka | IntTrajSim: Trajectory Prediction for Simulating Multi-Vehicle driving at Signalized Intersections | null | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Traffic simulators are widely used to study the operational efficiency of road infrastructure, but their rule-based approach limits their ability to mimic real-world driving behavior. Traffic intersections are critical components of the road infrastructure, both in terms of safety risk (nearly 28% of fatal crashes and 58% of nonfatal crashes happen at intersections) as well as the operational efficiency of a road corridor. This raises an important question: can we create a data-driven simulator that can mimic the macro- and micro-statistics of the driving behavior at a traffic intersection? Deep Generative Modeling-based trajectory prediction models provide a good starting point to model the complex dynamics of vehicles at an intersection. But they are not tested in a "live" micro-simulation scenario and are not evaluated on traffic engineering-related metrics. In this study, we propose traffic engineering-related metrics to evaluate generative trajectory prediction models and provide a simulation-in-the-loop pipeline to do so. We also provide a multi-headed self-attention-based trajectory prediction model that incorporates the signal information, which outperforms our previous models on the evaluation metrics. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:27:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Ranjan",
"Yash",
""
],
[
"Sengupta",
"Rahul",
""
],
[
"Rangarajan",
"Anand",
""
],
[
"Ranka",
"Sanjay",
""
]
] |
2506.08958 | Robert Style | Carla Fern\'andez-Rico and Robert W. Style and Stefanie Heyden and Shichen Wang and Peter D. Olmsted and Eric R. Dufresne | Thermodynamics of microphase separation in a swollen, strain-stiffening polymer network | null | null | null | null | cond-mat.soft cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Elastic MicroPhase Separation (EMPS) provides a simple route to create soft materials with homogeneous microstructures by leveraging the supersaturation of crosslinked polymer networks with liquids. At low supersaturation, network elasticity stabilizes a uniform mixture, but beyond a critical threshold, metastable microphase-separated domains emerge. While previous theories have focused on describing qualitative features about the size and morphology of these domains, they do not make quantitative predictions about EMPS phase diagrams. In this work, we extend Flory-Huggins theory to quantitatively capture EMPS phase diagrams by incorporating strain-stiffening effects. This model requires no fitting parameters and relies solely on independently measured solubility parameters and large-deformation mechanical responses. Our results reveal that strain-stiffening enables metastable microphase separation within the swelling equilibrium state and why the microstructures can range from discrete droplets to bicontinuous networks. This works highlights the critical role of nonlinear elasticity in controlling phase-separated morphologies in polymer gels. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:29:13 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Fernández-Rico",
"Carla",
""
],
[
"Style",
"Robert W.",
""
],
[
"Heyden",
"Stefanie",
""
],
[
"Wang",
"Shichen",
""
],
[
"Olmsted",
"Peter D.",
""
],
[
"Dufresne",
"Eric R.",
""
]
] |
2506.08959 | Paolo Arnaudo | Paolo Arnaudo, Giulio Bonelli, Alessandro Tanzini | One loop corrections to the thermodynamics of near-extremal Kerr-(A)dS black holes from Heun equation | 25 pages, 3 figures. Comments welcome | null | null | null | hep-th gr-qc math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | We compute one-loop corrections to the euclidean gravitational path integral of near-extremal (anti-)de Sitter-Kerr black hole in terms of the connection coefficients of the Heun equation describing the black hole linear perturbations in the Teukolsky formalism. We show that different near-extremal limits lead to distinct physical properties of the gravitational configuration, as they get described by distinct limiting differential equations. As a result, the light modes emerging in the limit determine different scaling properties in the temperature of the one-loop determinants. We show that the cold case displays distinctive universal log(T) corrections to the entropy of the system, including the ultracold regime. On the contrary, these do not appear in the limit in which the event horizon superimposes onto the cosmological one. In the Schwarzschild-de Sitter case, a further check is performed by comparison with the Denef-Hartnoll-Sachdev formula. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:30:52 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Arnaudo",
"Paolo",
""
],
[
"Bonelli",
"Giulio",
""
],
[
"Tanzini",
"Alessandro",
""
]
] |
2506.08961 | Chenxu Wang | Chenxu Wang and Huaping Liu | Towards Robust Deep Reinforcement Learning against Environmental State Perturbation | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Adversarial attacks and robustness in Deep Reinforcement Learning (DRL) have been widely studied in various threat models; however, few consider environmental state perturbations, which are natural in embodied scenarios. To improve the robustness of DRL agents, we formulate the problem of environmental state perturbation, introducing a preliminary non-targeted attack method as a calibration adversary, and then propose a defense framework, named Boosted Adversarial Training (BAT), which first tunes the agents via supervised learning to avoid catastrophic failure and subsequently adversarially trains the agent with reinforcement learning. Extensive experimental results substantiate the vulnerability of mainstream agents under environmental state perturbations and the effectiveness of our proposed attack. The defense results demonstrate that while existing robust reinforcement learning algorithms may not be suitable, our BAT framework can significantly enhance the robustness of agents against environmental state perturbations across various situations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:32:31 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Chenxu",
""
],
[
"Liu",
"Huaping",
""
]
] |
2506.08962 | Liangliang Chen | Liangliang Chen, Huiru Xie, Jacqueline Rohde, Ying Zhang | WIP: Large Language Model-Enhanced Smart Tutor for Undergraduate Circuit Analysis | Accepted to 2025 Frontiers in Education (FIE) Conference | null | null | null | cs.CY cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This research-to-practice work-in-progress (WIP) paper presents an AI-enabled smart tutor designed to provide homework assessment and feedback for students in an undergraduate circuit analysis course. We detail the tutor's design philosophy and core components, including open-ended question answering and homework feedback generation. The prompts are carefully crafted to optimize responses across different problems. The smart tutor was deployed on the Microsoft Azure platform and is currently in use in an undergraduate circuit analysis course at the School of Electrical and Computer Engineering in a large, public, research-intensive institution in the Southeastern United States. Beyond offering personalized instruction and feedback, the tutor collects student interaction data, which is summarized and shared with the course instructor. To evaluate its effectiveness, we collected student feedback, with 90.9% of responses indicating satisfaction with the tutor. Additionally, we analyze a subset of collected data on preliminary circuit analysis topics to assess tutor usage frequency for each problem and identify frequently asked questions. These insights help instructors gain real-time awareness of student difficulties, enabling more targeted classroom instruction. In future work, we will release a full analysis once the complete dataset is available after the Spring 2025 semester. We also explore the potential applications of this smart tutor across a broader range of engineering disciplines by developing improved prompts, diagram-recognition methods, and database management strategies, which remain ongoing areas of research. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:35:45 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Chen",
"Liangliang",
""
],
[
"Xie",
"Huiru",
""
],
[
"Rohde",
"Jacqueline",
""
],
[
"Zhang",
"Ying",
""
]
] |
2506.08963 | Yash Ranjan | Yash Ranjan, Rahul Sengupta, Anand Rangarajan, Sanjay Ranka | Evaluating Generative Vehicle Trajectory Models for Traffic Intersection Dynamics | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Traffic Intersections are vital to urban road networks as they regulate the movement of people and goods. However, they are regions of conflicting trajectories and are prone to accidents. Deep Generative models of traffic dynamics at signalized intersections can greatly help traffic authorities better understand the efficiency and safety aspects. At present, models are evaluated on computational metrics that primarily look at trajectory reconstruction errors. They are not evaluated online in a `live' microsimulation scenario. Further, these metrics do not adequately consider traffic engineering-specific concerns such as red-light violations, unallowed stoppage, etc. In this work, we provide a comprehensive analytics tool to train, run, and evaluate models with metrics that give better insights into model performance from a traffic engineering point of view. We train a state-of-the-art multi-vehicle trajectory forecasting model on a large dataset collected by running a calibrated scenario of a real-world urban intersection. We then evaluate the performance of the prediction models, online in a microsimulator, under unseen traffic conditions. We show that despite using ideally-behaved trajectories as input, and achieving low trajectory reconstruction errors, the generated trajectories show behaviors that break traffic rules. We introduce new metrics to evaluate such undesired behaviors and present our results. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:36:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Ranjan",
"Yash",
""
],
[
"Sengupta",
"Rahul",
""
],
[
"Rangarajan",
"Anand",
""
],
[
"Ranka",
"Sanjay",
""
]
] |
2506.08964 | Jinwoo Kim | Jinwoo Kim, Sangmin Han, Jinho Jeong, Jiwoo Choi, Dongyoung Kim, Seon Joo Kim | ORIDa: Object-centric Real-world Image Composition Dataset | Accepted at CVPR 2025 | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Object compositing, the task of placing and harmonizing objects in images of diverse visual scenes, has become an important task in computer vision with the rise of generative models. However, existing datasets lack the diversity and scale required to comprehensively explore real-world scenarios. We introduce ORIDa (Object-centric Real-world Image Composition Dataset), a large-scale, real-captured dataset containing over 30,000 images featuring 200 unique objects, each of which is presented across varied positions and scenes. ORIDa has two types of data: factual-counterfactual sets and factual-only scenes. The factual-counterfactual sets consist of four factual images showing an object in different positions within a scene and a single counterfactual (or background) image of the scene without the object, resulting in five images per scene. The factual-only scenes include a single image containing an object in a specific context, expanding the variety of environments. To our knowledge, ORIDa is the first publicly available dataset with its scale and complexity for real-world image composition. Extensive analysis and experiments highlight the value of ORIDa as a resource for advancing further research in object compositing. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:36:54 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Kim",
"Jinwoo",
""
],
[
"Han",
"Sangmin",
""
],
[
"Jeong",
"Jinho",
""
],
[
"Choi",
"Jiwoo",
""
],
[
"Kim",
"Dongyoung",
""
],
[
"Kim",
"Seon Joo",
""
]
] |
2506.08965 | Yiyang Zhao | Yiyang Zhao, Huiyu Bai, Xuejiao Zhao | GFRIEND: Generative Few-shot Reward Inference through EfficieNt DPO | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The ability to train high-performing reward models with few-shot data is critical for enhancing the efficiency and scalability of Reinforcement Learning from Human Feedback (RLHF). We propose a data augmentation and expansion framework that enables generative reward models trained on small datasets to achieve comparable performance to those trained on large-scale datasets. Traditional methods to train a generative reward model, such as Direct Preference Optimization (DPO), are constrained by inefficiencies in sample pairing and limited data diversity. This work introduces preference refinement, which employs Chain-of-Thought (CoT) sampling to uncover diverse and high-quality preference relationships. It also incorporates a perplexity-based scoring mechanism to assign nuanced preference levels and utilizes Multi-level Direct Preference Optimization (M-DPO) to enable the model to capture finer-grained preference differences between samples. Experimental results demonstrate that the proposed method significantly enhances data efficiency and model performance, enabling reward models trained in a few-shot setting to achieve results on par with those trained on large-scale datasets. This study underscores the potential of data-efficient strategies in advancing reward model optimization, offering a robust solution for low-resource RLHF applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:37:13 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhao",
"Yiyang",
""
],
[
"Bai",
"Huiyu",
""
],
[
"Zhao",
"Xuejiao",
""
]
] |
2506.08966 | Marek Kadl\v{c}\'ik | Marek Kadl\v{c}\'ik, Michal \v{S}tef\'anik, Timothee Mickus, Michal Spiegel, Josef Kucha\v{r} | Pre-trained Language Models Learn Remarkably Accurate Representations of Numbers | null | null | null | null | cs.CL cs.LG cs.NE | http://creativecommons.org/licenses/by/4.0/ | Pretrained language models (LMs) are prone to arithmetic errors. Existing work showed limited success in probing numeric values from models' representations, indicating that these errors can be attributed to the inherent unreliability of distributionally learned embeddings in representing exact quantities. However, we observe that previous probing methods are inadequate for the emergent structure of learned number embeddings with sinusoidal patterns.
In response, we propose a novel probing technique that decodes numeric values from input embeddings with near-perfect accuracy across a range of open-source LMs. This proves that after the sole pre-training, LMs represent numbers with remarkable precision. Finally, we find that the embeddings' preciseness judged by our probe's accuracy explains a large portion of LM's errors in elementary arithmetic, and show that aligning the embeddings with the pattern discovered by our probe can mitigate these errors. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:37:35 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Kadlčík",
"Marek",
""
],
[
"Štefánik",
"Michal",
""
],
[
"Mickus",
"Timothee",
""
],
[
"Spiegel",
"Michal",
""
],
[
"Kuchař",
"Josef",
""
]
] |
2506.08967 | Yuejiao Wang | Ailin Huang, Bingxin Li, Bruce Wang, Boyong Wu, Chao Yan, Chengli Feng, Heng Wang, Hongyu Zhou, Hongyuan Wang, Jingbei Li, Jianjian Sun, Joanna Wang, Mingrui Chen, Peng Liu, Ruihang Miao, Shilei Jiang, Tian Fei, Wang You, Xi Chen, Xuerui Yang, Yechang Huang, Yuxiang Zhang, Zheng Ge, Zheng Gong, Zhewei Huang, Zixin Zhang, Bin Wang, Bo Li, Buyun Ma, Changxin Miao, Changyi Wan, Chen Xu, Dapeng Shi, Dingyuan Hu, Enle Liu, Guanzhe Huang, Gulin Yan, Hanpeng Hu, Haonan Jia, Jiahao Gong, Jiaoren Wu, Jie Wu, Jie Yang, Junzhe Lin, Kaixiang Li, Lei Xia, Longlong Gu, Ming Li, Nie Hao, Ranchen Ming, Shaoliang Pang, Siqi Liu, Song Yuan, Tiancheng Cao, Wen Li, Wenqing He, Xu Zhao, Xuelin Zhang, Yanbo Yu, Yinmin Zhong, Yu Zhou, Yuanwei Liang, Yuanwei Lu, Yuxiang Yang, Zidong Yang, Zili Zhang, Binxing Jiao, Heung-Yeung Shum, Jiansheng Chen, Jing Li, Xiangyu Zhang, Xinhao Zhang, Yibo Zhu, Daxin Jiang, Shuchang Zhou, Chen Hu | Step-Audio-AQAA: a Fully End-to-End Expressive Large Audio Language Model | 12 pages, 3 figures | null | null | null | cs.SD cs.CL eess.AS | http://creativecommons.org/licenses/by-sa/4.0/ | Large Audio-Language Models (LALMs) have significantly advanced intelligent human-computer interaction, yet their reliance on text-based outputs limits their ability to generate natural speech responses directly, hindering seamless audio interactions. To address this, we introduce Step-Audio-AQAA, a fully end-to-end LALM designed for Audio Query-Audio Answer (AQAA) tasks. The model integrates a dual-codebook audio tokenizer for linguistic and semantic feature extraction, a 130-billion-parameter backbone LLM and a neural vocoder for high-fidelity speech synthesis. Our post-training approach employs interleaved token-output of text and audio to enhance semantic coherence and combines Direct Preference Optimization (DPO) with model merge to improve performance. Evaluations on the StepEval-Audio-360 benchmark demonstrate that Step-Audio-AQAA excels especially in speech control, outperforming the state-of-art LALMs in key areas. This work contributes a promising solution for end-to-end LALMs and highlights the critical role of token-based vocoder in enhancing overall performance for AQAA tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:37:39 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Huang",
"Ailin",
""
],
[
"Li",
"Bingxin",
""
],
[
"Wang",
"Bruce",
""
],
[
"Wu",
"Boyong",
""
],
[
"Yan",
"Chao",
""
],
[
"Feng",
"Chengli",
""
],
[
"Wang",
"Heng",
""
],
[
"Zhou",
"Hongyu",
""
],
[
"Wang",
"Hongyuan",
""
],
[
"Li",
"Jingbei",
""
],
[
"Sun",
"Jianjian",
""
],
[
"Wang",
"Joanna",
""
],
[
"Chen",
"Mingrui",
""
],
[
"Liu",
"Peng",
""
],
[
"Miao",
"Ruihang",
""
],
[
"Jiang",
"Shilei",
""
],
[
"Fei",
"Tian",
""
],
[
"You",
"Wang",
""
],
[
"Chen",
"Xi",
""
],
[
"Yang",
"Xuerui",
""
],
[
"Huang",
"Yechang",
""
],
[
"Zhang",
"Yuxiang",
""
],
[
"Ge",
"Zheng",
""
],
[
"Gong",
"Zheng",
""
],
[
"Huang",
"Zhewei",
""
],
[
"Zhang",
"Zixin",
""
],
[
"Wang",
"Bin",
""
],
[
"Li",
"Bo",
""
],
[
"Ma",
"Buyun",
""
],
[
"Miao",
"Changxin",
""
],
[
"Wan",
"Changyi",
""
],
[
"Xu",
"Chen",
""
],
[
"Shi",
"Dapeng",
""
],
[
"Hu",
"Dingyuan",
""
],
[
"Liu",
"Enle",
""
],
[
"Huang",
"Guanzhe",
""
],
[
"Yan",
"Gulin",
""
],
[
"Hu",
"Hanpeng",
""
],
[
"Jia",
"Haonan",
""
],
[
"Gong",
"Jiahao",
""
],
[
"Wu",
"Jiaoren",
""
],
[
"Wu",
"Jie",
""
],
[
"Yang",
"Jie",
""
],
[
"Lin",
"Junzhe",
""
],
[
"Li",
"Kaixiang",
""
],
[
"Xia",
"Lei",
""
],
[
"Gu",
"Longlong",
""
],
[
"Li",
"Ming",
""
],
[
"Hao",
"Nie",
""
],
[
"Ming",
"Ranchen",
""
],
[
"Pang",
"Shaoliang",
""
],
[
"Liu",
"Siqi",
""
],
[
"Yuan",
"Song",
""
],
[
"Cao",
"Tiancheng",
""
],
[
"Li",
"Wen",
""
],
[
"He",
"Wenqing",
""
],
[
"Zhao",
"Xu",
""
],
[
"Zhang",
"Xuelin",
""
],
[
"Yu",
"Yanbo",
""
],
[
"Zhong",
"Yinmin",
""
],
[
"Zhou",
"Yu",
""
],
[
"Liang",
"Yuanwei",
""
],
[
"Lu",
"Yuanwei",
""
],
[
"Yang",
"Yuxiang",
""
],
[
"Yang",
"Zidong",
""
],
[
"Zhang",
"Zili",
""
],
[
"Jiao",
"Binxing",
""
],
[
"Shum",
"Heung-Yeung",
""
],
[
"Chen",
"Jiansheng",
""
],
[
"Li",
"Jing",
""
],
[
"Zhang",
"Xiangyu",
""
],
[
"Zhang",
"Xinhao",
""
],
[
"Zhu",
"Yibo",
""
],
[
"Jiang",
"Daxin",
""
],
[
"Zhou",
"Shuchang",
""
],
[
"Hu",
"Chen",
""
]
] |
2506.08968 | Amirreza Rouhi | Amirreza Rouhi, Solmaz Arezoomandan, Knut Peterson, Joseph T. Woods, David K. Han | ADAM: Autonomous Discovery and Annotation Model using LLMs for Context-Aware Annotations | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Object detection models typically rely on predefined categories, limiting their ability to identify novel objects in open-world scenarios. To overcome this constraint, we introduce ADAM: Autonomous Discovery and Annotation Model, a training-free, self-refining framework for open-world object labeling. ADAM leverages large language models (LLMs) to generate candidate labels for unknown objects based on contextual information from known entities within a scene. These labels are paired with visual embeddings from CLIP to construct an Embedding-Label Repository (ELR) that enables inference without category supervision. For a newly encountered unknown object, ADAM retrieves visually similar instances from the ELR and applies frequency-based voting and cross-modal re-ranking to assign a robust label. To further enhance consistency, we introduce a self-refinement loop that re-evaluates repository labels using visual cohesion analysis and k-nearest-neighbor-based majority re-labeling. Experimental results on the COCO and PASCAL datasets demonstrate that ADAM effectively annotates novel categories using only visual and contextual signals, without requiring any fine-tuning or retraining. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:41:33 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Rouhi",
"Amirreza",
""
],
[
"Arezoomandan",
"Solmaz",
""
],
[
"Peterson",
"Knut",
""
],
[
"Woods",
"Joseph T.",
""
],
[
"Han",
"David K.",
""
]
] |
2506.08969 | Kevin Luhman | K. L. Luhman, C. Alves de Oliveira | A New Spectral Class of Brown Dwarfs at the Bottom of the IMF in IC 348 | The Astrophysical Journal Letters, in press. A description of these results for the public is available at https://sites.psu.edu/kluhman/ic348/ | null | null | null | astro-ph.GA astro-ph.EP astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | In a previous study, we used JWST to identify three new brown dwarfs in the center of a nearby star-forming cluster, IC 348. The faintest object had an estimated mass of 3-4 $M_{\rm Jup}$, making it a contender for the least massive brown dwarf confirmed with spectroscopy. Two of the new brown dwarfs also exhibited absorption features from an unidentified aliphatic hydrocarbon, which were not predicted by atmospheric models and were not previously detected in atmospheres outside of the solar system. We have used JWST to perform a deeper survey for brown dwarfs across a larger field in IC 348. We have identified 39 brown dwarf candidates in NIRCam images and have obtained spectra for 15 of them with NIRSpec, nine of which are classified as substellar members of the cluster. The faintest new members have mass estimates of $\sim2$ $M_{\rm Jup}$, providing a new constraint on the minimum mass of the IMF. Two new members ($\sim2$ and 10 $M_{\rm Jup}$) exhibit large excess emission from circumstellar disks, demonstrating that they harbor the raw materials for planet formation. Finally, eight of the nine new brown dwarfs and one known member that is newly observed with NIRSpec show the aforementioned hydrocarbon features. Among the total of 11 brown dwarfs in IC 348 that have hydrocarbon detections, the features are stronger at fainter magnitudes, indicating that the hydrocarbon is a natural constituent of the atmospheres of the coolest newborn brown dwarfs. We propose a new spectral class "H" that is defined by the presence of the 3.4 $\mu$m fundamental band of the hydrocarbon. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:42:48 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Luhman",
"K. L.",
""
],
[
"de Oliveira",
"C. Alves",
""
]
] |
2506.08970 | Wei Jiyao | Jiyao Wei, Saiping Guan, Da Li, Xiaolong Jin, Jiafeng Guo, Xueqi Cheng | A Survey of Link Prediction in N-ary Knowledge Graphs | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | N-ary Knowledge Graphs (NKGs) are a specialized type of knowledge graph designed to efficiently represent complex real-world facts. Unlike traditional knowledge graphs, where a fact typically involves two entities, NKGs can capture n-ary facts containing more than two entities. Link prediction in NKGs aims to predict missing elements within these n-ary facts, which is essential for completing NKGs and improving the performance of downstream applications. This task has recently gained significant attention. In this paper, we present the first comprehensive survey of link prediction in NKGs, providing an overview of the field, systematically categorizing existing methods, and analyzing their performance and application scenarios. We also outline promising directions for future research. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:44:27 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wei",
"Jiyao",
""
],
[
"Guan",
"Saiping",
""
],
[
"Li",
"Da",
""
],
[
"Jin",
"Xiaolong",
""
],
[
"Guo",
"Jiafeng",
""
],
[
"Cheng",
"Xueqi",
""
]
] |
2506.08971 | Mat\'ias Rub\'en Bola\~nos | Kannan Vijayadharan, Mat\'ias Rub\'en Bola\~nos, Marco Avesani, Giuseppe Vallone, Paolo Villoresi, Costantino Agnesi | A Sagnac-based arbitrary time-bin state encoder for quantum communication applications | 10 pages, 12 figures | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Time-bin encoding of quantum information is highly advantageous for long-distance quantum communication protocols over optical fibres due to its inherent robustness in the channel and the possibility of generating high-dimensional quantum states. The most common implementation of time-bin quantum states using unbalanced interferometers presents challenges in terms of stability and flexibility of operation. In particular, a limited number of states can be generated without modifying the optical scheme. Here we present the implementation of a fully controllable arbitrary time-bin quantum state encoder, which is easily scalable to arbitrary dimensions and time-bin widths. The encoder presents high stability and low quantum bit error rate QBER, even at high speeds of operation. Additionally, we demonstrate phase randomization and phase encoding without additional resources. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:44:47 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Vijayadharan",
"Kannan",
""
],
[
"Bolaños",
"Matías Rubén",
""
],
[
"Avesani",
"Marco",
""
],
[
"Vallone",
"Giuseppe",
""
],
[
"Villoresi",
"Paolo",
""
],
[
"Agnesi",
"Costantino",
""
]
] |
2506.08972 | Yuan Guo | Yuan Guo, Tingjia Miao, Zheng Wu, Pengzhou Cheng, Ming Zhou, Zhuosheng Zhang | Atomic-to-Compositional Generalization for Mobile Agents with A New Benchmark and Scheduling System | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous agents powered by multimodal large language models have been developed to facilitate task execution on mobile devices. However, prior work has predominantly focused on atomic tasks -- such as shot-chain execution tasks and single-screen grounding tasks -- while overlooking the generalization to compositional tasks, which are indispensable for real-world applications. This work introduces UI-NEXUS, a comprehensive benchmark designed to evaluate mobile agents on three categories of compositional operations: Simple Concatenation, Context Transition, and Deep Dive. UI-NEXUS supports interactive evaluation in 20 fully controllable local utility app environments, as well as 30 online Chinese and English service apps. It comprises 100 interactive task templates with an average optimal step count of 14.05. Experimental results across a range of mobile agents with agentic workflow or agent-as-a-model show that UI-NEXUS presents significant challenges. Specifically, existing agents generally struggle to balance performance and efficiency, exhibiting representative failure modes such as under-execution, over-execution, and attention drift, causing visible atomic-to-compositional generalization gap. Inspired by these findings, we propose AGENT-NEXUS, a lightweight and efficient scheduling system to tackle compositional mobile tasks. AGENT-NEXUS extrapolates the abilities of existing mobile agents by dynamically decomposing long-horizon tasks to a series of self-contained atomic subtasks. AGENT-NEXUS achieves 24% to 40% task success rate improvement for existing mobile agents on compositional operation tasks within the UI-NEXUS benchmark without significantly sacrificing inference overhead. The demo video, dataset, and code are available on the project page at https://ui-nexus.github.io. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:45:29 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Guo",
"Yuan",
""
],
[
"Miao",
"Tingjia",
""
],
[
"Wu",
"Zheng",
""
],
[
"Cheng",
"Pengzhou",
""
],
[
"Zhou",
"Ming",
""
],
[
"Zhang",
"Zhuosheng",
""
]
] |
2506.08973 | Vladimir Rovenski | Vladimir Rovenski | $\ast$-$\eta$-Ricci solitons on weak Kenmotsu $f$-manifolds | 18 pages. arXiv admin note: text overlap with arXiv:2503.18031, arXiv:2412.14125 | null | null | null | math.DG | http://creativecommons.org/licenses/by/4.0/ | Recent interest among geometers in $f$-structures of K. Yano is due to the study of topology and dynamics of contact foliations and generalized A. Weinstein conjectures. Weak metric $f$-structures, introduced by the author and R. Wolak as a generalization of Hermitian structure, as well as $f$-structure allow for a fresh perspective on the classical theory. An important case of such manifolds, which is locally a twisted product, is a weak $\beta f$-Kenmotsu manifold defined as a generalization of K. Kenmotsu's concept. In this paper, the concept of the $\ast$-Ricci tensor of S. Tashibana is adapted to weak metric $f$-manifolds, the interaction of $\ast$-$\eta$-Ricci soliton with the weak $\beta f$-Kenmotsu structure is studied and new characteristics of $\eta$-Einstein metrics are obtained. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:45:41 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Rovenski",
"Vladimir",
""
]
] |
2506.08974 | Igor Kvasic | Igor Kvasi\'c, Derek Orbaugh Antillon, {\DH}ula Na{\dj}, Christopher Walker, Iain Anderson, Nikola Mi\v{s}kovi\'c | Diver-Robot Communication Dataset for Underwater Hand Gesture Recognition | 28 pages, 12 figures (Elsevier Computer Networks Journal) | Computer Networks, Volume 245, May 2024, 110392 | 10.1016/j.comnet.2024.110392 | null | eess.IV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we present a dataset of diving gesture images used for human-robot interaction underwater. By offering this open access dataset, the paper aims at investigating the potential of using visual detection of diving gestures from an autonomous underwater vehicle (AUV) as a form of communication with a human diver. In addition to the image recording, the same dataset was recorded using a smart gesture recognition glove. The glove uses elastomer sensors and on-board processing to determine the selected gesture and transmit the command associated with the gesture to the AUV via acoustics. Although this method can be used under different visibility conditions and even without line of sight, it introduces a communication delay required for the acoustic transmission of the gesture command. To compare efficiency, the glove was equipped with visual markers proposed in a gesture-based language called CADDIAN and recorded with an underwater camera in parallel to the glove's onboard recognition process. The dataset contains over 30,000 underwater frames of nearly 900 individual gestures annotated in corresponding snippet folders. The dataset was recorded in a balanced ratio with five different divers in sea and five different divers in pool conditions, with gestures recorded at 1, 2 and 3 metres from the camera. The glove gesture recognition statistics are reported in terms of average diver reaction time, average time taken to perform a gesture, recognition success rate, transmission times and more. The dataset presented should provide a good baseline for comparing the performance of state of the art visual diving gesture recognition techniques under different visibility conditions. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:45:45 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Kvasić",
"Igor",
""
],
[
"Antillon",
"Derek Orbaugh",
""
],
[
"Nađ",
"Ðula",
""
],
[
"Walker",
"Christopher",
""
],
[
"Anderson",
"Iain",
""
],
[
"Mišković",
"Nikola",
""
]
] |
2506.08975 | Alireza Fereidunian | Masoud Behbahani and Alireza Fereidunian | Quantitative Indices for Improving Metro Load Curve, Using Distributed Generation | Preprint of the accepted paper for the EPDC 2014 | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper promises the idea of using DG (Distributed Generation) to improve the Metro load curve. Public transportation systems are often based on gasoline and diesel. However, with the gradual development in usage of the Metro and monorail, a new load with heavy demand, inappropriate load curve and middle LF (Load factor) is added to the electricity grid. In addition to supply problem of this massive consumer, the Metro load curve is another problem, which has a relatively low LF. Furthermore, Metro load peak hours coincide with the peaks of national grid. Improvement of the load curve is well-known in electrical engineering literature, which depending on the type of load curve, offers general recommendations in three approaches; DSM (Demand Side Management), DS (Distributed Storage) and DG. In this paper, to achieve quantitative indices of improvement for Metro load curve using DG, firstly based on the analysis of volume and consumption pattern of the main loads in Metro, the typical load curve has been extracted. Using this curve, the result of using DG is shown by quantitative parameters which represent the significant improvement in load curve. These parameters can be used to calculate economic indicators such as initial cost and ROI (Return of Investment). | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:45:49 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Behbahani",
"Masoud",
""
],
[
"Fereidunian",
"Alireza",
""
]
] |
2506.08977 | Victoria Hankemeier | Victoria Hankemeier, Malte Schilling | Tailored Architectures for Time Series Forecasting: Evaluating Deep Learning Models on Gaussian Process-Generated Data | Accepted at IJCNN25, Code: https://github.com/vicky-hnk/time-flex | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Developments in Deep Learning have significantly improved time series forecasting by enabling more accurate modeling of complex temporal dependencies inherent in sequential data. The effectiveness of such models is often demonstrated on limited sets of specific real-world data. Although this allows for comparative analysis, it still does not demonstrate how specific data characteristics align with the architectural strengths of individual models. Our research aims at uncovering clear connections between time series characteristics and particular models. We introduce a novel dataset generated using Gaussian Processes, specifically designed to display distinct, known characteristics for targeted evaluations of model adaptability to them. Furthermore, we present TimeFlex, a new model that incorporates a modular architecture tailored to handle diverse temporal dynamics, including trends and periodic patterns. This model is compared to current state-of-the-art models, offering a deeper understanding of how models perform under varied time series conditions. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:46:02 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Hankemeier",
"Victoria",
""
],
[
"Schilling",
"Malte",
""
]
] |
2506.08978 | Willem Zuidema | Anna Langedijk, Jaap Jumelet, Willem Zuidema | Propositional Logic for Probing Generalization in Neural Networks | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | The extent to which neural networks are able to acquire and represent symbolic rules remains a key topic of research and debate. Much current work focuses on the impressive capabilities of large language models, as well as their often ill-understood failures on a wide range of reasoning tasks. In this paper, in contrast, we investigate the generalization behavior of three key neural architectures (Transformers, Graph Convolution Networks and LSTMs) in a controlled task rooted in propositional logic. The task requires models to generate satisfying assignments for logical formulas, making it a structured and interpretable setting for studying compositionality. We introduce a balanced extension of an existing dataset to eliminate superficial patterns and enable testing on unseen operator combinations. Using this dataset, we evaluate the ability of the three architectures to generalize beyond the training distribution. While all models perform well in-distribution, we find that generalization to unseen patterns, particularly those involving negation, remains a significant challenge. Transformers fail to apply negation compositionally, unless structural biases are introduced. Our findings highlight persistent limitations in the ability of standard architectures to learn systematic representations of logical operators, suggesting the need for stronger inductive biases to support robust rule-based reasoning. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:46:05 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Langedijk",
"Anna",
""
],
[
"Jumelet",
"Jaap",
""
],
[
"Zuidema",
"Willem",
""
]
] |
2506.08979 | Ping Hu | Longyu Yang, Ping Hu, Lu Zhang, Jun Liu, Yap-Peng Tan, Heng Tao Shen, Xiaofeng Zhu | Rethinking Range-View LiDAR Segmentation in Adverse Weather | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | LiDAR segmentation has emerged as an important task to enrich multimedia experiences and analysis. Range-view-based methods have gained popularity due to their high computational efficiency and compatibility with real-time deployment. However, their generalized performance under adverse weather conditions remains underexplored, limiting their reliability in real-world environments. In this work, we identify and analyze the unique challenges that affect the generalization of range-view LiDAR segmentation in severe weather. To address these challenges, we propose a modular and lightweight framework that enhances robustness without altering the core architecture of existing models. Our method reformulates the initial stem block of standard range-view networks into two branches to process geometric attributes and reflectance intensity separately. Specifically, a Geometric Abnormality Suppression (GAS) module reduces the influence of weather-induced spatial noise, and a Reflectance Distortion Calibration (RDC) module corrects reflectance distortions through memory-guided adaptive instance normalization. The processed features are then fused and passed to the original segmentation pipeline. Extensive experiments on different benchmarks and baseline models demonstrate that our approach significantly improves generalization to adverse weather with minimal inference overhead, offering a practical and effective solution for real-world LiDAR segmentation. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:48:27 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Yang",
"Longyu",
""
],
[
"Hu",
"Ping",
""
],
[
"Zhang",
"Lu",
""
],
[
"Liu",
"Jun",
""
],
[
"Tan",
"Yap-Peng",
""
],
[
"Shen",
"Heng Tao",
""
],
[
"Zhu",
"Xiaofeng",
""
]
] |
2506.08981 | Satu Hopponen | Satu Hopponen, Tomi Kinnunen, Alexandre Nikolaev, Rosa Gonz\'alez Hautam\"aki, Lauri Tavi, Einar Meister | FROST-EMA: Finnish and Russian Oral Speech Dataset of Electromagnetic Articulography Measurements with L1, L2 and Imitated L2 Accents | Accepted in Interspeech 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | We introduce a new FROST-EMA (Finnish and Russian Oral Speech Dataset of Electromagnetic Articulography) corpus. It consists of 18 bilingual speakers, who produced speech in their native language (L1), second language (L2), and imitated L2 (fake foreign accent). The new corpus enables research into language variability from phonetic and technological points of view. Accordingly, we include two preliminary case studies to demonstrate both perspectives. The first case study explores the impact of L2 and imitated L2 on the performance of an automatic speaker verification system, while the second illustrates the articulatory patterns of one speaker in L1, L2, and a fake accent. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:52:11 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Hopponen",
"Satu",
""
],
[
"Kinnunen",
"Tomi",
""
],
[
"Nikolaev",
"Alexandre",
""
],
[
"Hautamäki",
"Rosa González",
""
],
[
"Tavi",
"Lauri",
""
],
[
"Meister",
"Einar",
""
]
] |
2506.08983 | Yue Wu | Yue Wu, Jianfu Cao, Ye Cao | Online Learning Control Strategies for Industrial Processes with Application for Loosening and Conditioning | 19pages,6figures | null | null | null | eess.SY cs.SY math.OA | http://creativecommons.org/licenses/by/4.0/ | This paper proposes a novel adaptive Koopman Model Predictive Control (MPC) framework, termed HPC-AK-MPC, designed to address the dual challenges of time-varying dynamics and safe operation in complex industrial processes. The framework integrates two core strategies: online learning and historically-informed safety constraints. To contend with process time-variance, a Recursive Extended Dynamic Mode Decomposition (rEDMDc) technique is employed to construct an adaptive Koopman model capable of updating its parameters from real-time data, endowing the controller with the ability to continuously learn and track dynamic changes. To tackle the critical issue of safe operation under model uncertainty, we introduce a novel Historical Process Constraint (HPC) mechanism. This mechanism mines successful operational experiences from a historical database and, by coupling them with the confidence level of the online model, generates a dynamic "safety corridor" for the MPC optimization problem. This approach transforms implicit expert knowledge into explicit, adaptive constraints, establishing a dynamic balance between pursuing optimal performance and ensuring robust safety. The proposed HPC-AK-MPC method is applied to a real-world tobacco loosening and conditioning process and systematically validated using an "advisor mode" simulation framework with industrial data. Experimental results demonstrate that, compared to historical operations, the proposed method significantly improves the Process Capability Index (Cpk) for key quality variables across all tested batches, proving its substantial potential in enhancing control performance while guaranteeing operational safety. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:53:00 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wu",
"Yue",
""
],
[
"Cao",
"Jianfu",
""
],
[
"Cao",
"Ye",
""
]
] |
2506.08984 | Valeska Zambra | Valeska Zambra, Amit Nathwani, Muhammad Nauman, Sylvia K. Lewin, Corey E. Frank, Nicholas P. Butch, Arkady Shekhter, B. J. Ramshaw, K. A. Modic | Giant transverse magnetic fluctuations at the edge of re-entrant superconductivity in UTe$_{2}$ | null | null | null | null | cond-mat.str-el cond-mat.supr-con | http://creativecommons.org/licenses/by/4.0/ | UTe$_{2}$ exhibits the remarkable phenomenon of re-entrant superconductivity, whereby the zero-resistance state reappears above 40 tesla after being suppressed with a field of around 10 tesla. One potential pairing mechanism, invoked in the related re-entrant superconductors UCoGe and URhGe, involves transverse fluctuations of a ferromagnetic order parameter. However, the requisite ferromagnetic order - present in both UCoGe and URhGe - is absent in UTe$_{2}$, and magnetization measurements show no sign of strong fluctuations. Here, we measure the magnetotropic susceptibility of UTe$_{2}$ across two field-angle planes. This quantity is sensitive to the magnetic susceptibility in a direction transverse to the applied magnetic field - a quantity that is not accessed in conventional magnetization measurements. We observe a very large decrease in the magnetotropic susceptibility over a broad range of field orientations, indicating a large increase in the transverse magnetic susceptibility. The three superconducting phases of UTe$_{2}$, including the high-field re-entrant phase, surround this region of enhanced susceptibility in the field-angle phase diagram. The strongest transverse susceptibility is found near the critical end point of the high-field metamagnetic transition, suggesting that quantum critical fluctuations of a field-induced magnetic order parameter may be responsible for the large transverse susceptibility, and may provide a pairing mechanism for field-induced superconductivity in UTe$_{2}$. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:57:08 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zambra",
"Valeska",
""
],
[
"Nathwani",
"Amit",
""
],
[
"Nauman",
"Muhammad",
""
],
[
"Lewin",
"Sylvia K.",
""
],
[
"Frank",
"Corey E.",
""
],
[
"Butch",
"Nicholas P.",
""
],
[
"Shekhter",
"Arkady",
""
],
[
"Ramshaw",
"B. J.",
""
],
[
"Modic",
"K. A.",
""
]
] |
2506.08985 | Venkat Chandrasekhar | Venkat Chandrasekhar | Tuning the the fundamental periodicity of the current-phase relation in multiterminal diffusive Josephson junctions | null | null | null | null | cond-mat.supr-con quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Conventional superconductor/insulator/superconductor (SIS) Josephson junctions, devices where two superconductors are separated by a tunnel barrier are technologically important as elements in quantum circuits, particularly with their key role in superconducting qubits. An important characteristic of Josephson junctions is the relation between the supercurrent Is and the phase difference {\phi} between them. For SIS junctions, the current-phase relation is sinusioidal and 2{\pi} periodic. Other types of Josephson junctions, where the material between the superconductors is a weak link or a normal metal (N) may have non-sinusoidal current-phase relations that are still 2{\pi} periodic. We show here that a multi-terminal diffusive SNS Josephson junction with 4 superconducting contacts can show a current phase relation between two of the contacts that is a superposition of 2{\pi} and 4{\pi} periodic components whose relative strength is controlled by the phase difference between the other two contacts, becoming 2{\pi} or 4{\pi} periodic for certain values of this phase difference. This tunability might have applications in tailoring the Hamiltonians of superconducting quantum circuits. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:58:32 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Chandrasekhar",
"Venkat",
""
]
] |
2506.08986 | Yuejiao Wang | Yuejiao Wang, Xianmin Gong, Xixin Wu, Patrick Wong, Hoi-lam Helene Fung, Man Wai Mak, Helen Meng | Naturalistic Language-related Movie-Watching fMRI Task for Detecting Neurocognitive Decline and Disorder | 5 pages,3 figures, accepted by ISCSLP 2024 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Early detection is crucial for timely intervention aimed at preventing and slowing the progression of neurocognitive disorder (NCD), a common and significant health problem among the aging population. Recent evidence has suggested that language-related functional magnetic resonance imaging (fMRI) may be a promising approach for detecting cognitive decline and early NCD. In this paper, we proposed a novel, naturalistic language-related fMRI task for this purpose. We examined the effectiveness of this task among 97 non-demented Chinese older adults from Hong Kong. The results showed that machine-learning classification models based on fMRI features extracted from the task and demographics (age, gender, and education year) achieved an average area under the curve of 0.86 when classifying participants' cognitive status (labeled as NORMAL vs DECLINE based on their scores on a standard neurcognitive test). Feature localization revealed that the fMRI features most frequently selected by the data-driven approach came primarily from brain regions associated with language processing, such as the superior temporal gyrus, middle temporal gyrus, and right cerebellum. The study demonstrated the potential of the naturalistic language-related fMRI task for early detection of aging-related cognitive decline and NCD. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 16:58:47 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Yuejiao",
""
],
[
"Gong",
"Xianmin",
""
],
[
"Wu",
"Xixin",
""
],
[
"Wong",
"Patrick",
""
],
[
"Fung",
"Hoi-lam Helene",
""
],
[
"Mak",
"Man Wai",
""
],
[
"Meng",
"Helen",
""
]
] |
2506.08987 | Ehsan Naghavi | Ehsan Naghavi, Haifeng Wang, Vahid Ziaei Rad, Julius Guccione, Ghassan Kassab, Vishnu Boddeti, Seungik Baek, Lik-Chuan Lee | Rapid cardiac activation prediction for cardiac resynchronization therapy planning using geometric deep learning | null | null | null | null | cs.CE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cardiac resynchronization therapy (CRT) is a common intervention for patients with dyssynchronous heart failure, yet approximately one-third of recipients fail to respond due to suboptimal lead placement. Identifying optimal pacing sites remains challenging, largely due to patient-specific anatomical variability and the limitations of current individualized planning strategies. In a step towards constructing an in-silico approach to help address this issue, we develop two geometric deep learning (DL) models, based on graph neural network (GNN) and geometry-informed neural operator (GINO), to predict cardiac activation time map in real-time for CRT planning and optimization. Both models are trained on a large synthetic dataset generated from finite-element (FE) simulations over a wide range of left ventricular (LV) geometries, pacing site configurations, and tissue conductivities. The GINO model significantly outperforms the GNN model, with lower prediction errors (1.14% vs 3.14%) and superior robustness to noise and various mesh discretization. Using the GINO model, we also develop a workflow for optimizing the pacing site in CRT from given activation time map and LV geometry. Compared to randomly selecting a pacing site, the CRT optimization workflow produces a larger reduction in maximum activation time (20% vs. 8%). In conjunction with an interactive web-based graphical user interface (GUI) available at https://dcsim.egr.msu.edu/, the GINO model shows promising potential as a clinical decision-support tool for personalized pre-procedural CRT optimization. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:00:31 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Naghavi",
"Ehsan",
""
],
[
"Wang",
"Haifeng",
""
],
[
"Rad",
"Vahid Ziaei",
""
],
[
"Guccione",
"Julius",
""
],
[
"Kassab",
"Ghassan",
""
],
[
"Boddeti",
"Vishnu",
""
],
[
"Baek",
"Seungik",
""
],
[
"Lee",
"Lik-Chuan",
""
]
] |
2506.08988 | Youyicun Lin | Liang Chen, Youyicun Lin, Yuxuan Zhou | An Efficient Augmented Lagrangian Method for Dynamic Optimal Transport on Surfaces Based on Second-Order Cone Programming | null | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes an efficient numerical optimization approach for solving dynamic optimal transport (DOT) problems on general smooth surfaces, computing both the quadratic Wasserstein distance and the associated transportation path. Building on the convex DOT model of Benamou and Brenier, we first properly reformulate its dual problem, discretized on a triangular mesh for space together with a staggered grid for time, to a linear second-order cone programming. Then the resulting finite-dimensional convex optimization problem is solved via an inexact semi-proximal augmented Lagrangian method with a highly efficient numerical implementation, and the algorithm is guaranteed to converge to a Karush-Kuhn-Tucker solution without imposing any additional assumptions. Finally, we implement the proposed methodology as an open-source software package. The effectiveness, robustness, and computational efficiency of the software are demonstrated through extensive numerical experiments across diverse datasets, where it consistently outperforms state-of-the-art solvers by several times in speed. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:01:28 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Chen",
"Liang",
""
],
[
"Lin",
"Youyicun",
""
],
[
"Zhou",
"Yuxuan",
""
]
] |
2506.08989 | Xiao Liang | Xiao Liang, Zhong-Zhi Li, Yeyun Gong, Yang Wang, Hengyuan Zhang, Yelong Shen, Ying Nian Wu, Weizhu Chen | SwS: Self-aware Weakness-driven Problem Synthesis in Reinforcement Learning for LLM Reasoning | Reinforcement Learning; Large Language Models; LLM Reasoning | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | Reinforcement Learning with Verifiable Rewards (RLVR) has proven effective for training large language models (LLMs) on complex reasoning tasks, such as mathematical problem solving. A prerequisite for the scalability of RLVR is a high-quality problem set with precise and verifiable answers. However, the scarcity of well-crafted human-labeled math problems and limited-verification answers in existing distillation-oriented synthetic datasets limit their effectiveness in RL. Additionally, most problem synthesis strategies indiscriminately expand the problem set without considering the model's capabilities, leading to low efficiency in generating useful questions. To mitigate this issue, we introduce a Self-aware Weakness-driven problem Synthesis framework (SwS) that systematically identifies model deficiencies and leverages them for problem augmentation. Specifically, we define weaknesses as questions that the model consistently fails to learn through its iterative sampling during RL training. We then extract the core concepts from these failure cases and synthesize new problems to strengthen the model's weak areas in subsequent augmented training, enabling it to focus on and gradually overcome its weaknesses. Without relying on external knowledge distillation, our framework enables robust generalization byempowering the model to self-identify and address its weaknesses in RL, yielding average performance gains of 10.0% and 7.7% on 7B and 32B models across eight mainstream reasoning benchmarks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:02:00 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Liang",
"Xiao",
""
],
[
"Li",
"Zhong-Zhi",
""
],
[
"Gong",
"Yeyun",
""
],
[
"Wang",
"Yang",
""
],
[
"Zhang",
"Hengyuan",
""
],
[
"Shen",
"Yelong",
""
],
[
"Wu",
"Ying Nian",
""
],
[
"Chen",
"Weizhu",
""
]
] |
2506.08990 | Chenyu Lian | Chenyu Lian, Hong-Yu Zhou, Dongyun Liang, Jing Qin, Liansheng Wang | Efficient Medical Vision-Language Alignment Through Adapting Masked Vision Models | TMI 2025 | null | 10.1109/TMI.2025.3575853 | null | cs.CV cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Medical vision-language alignment through cross-modal contrastive learning shows promising performance in image-text matching tasks, such as retrieval and zero-shot classification. However, conventional cross-modal contrastive learning (CLIP-based) methods suffer from suboptimal visual representation capabilities, which also limits their effectiveness in vision-language alignment. In contrast, although the models pretrained via multimodal masked modeling struggle with direct cross-modal matching, they excel in visual representation. To address this contradiction, we propose ALTA (ALign Through Adapting), an efficient medical vision-language alignment method that utilizes only about 8% of the trainable parameters and less than 1/5 of the computational consumption required for masked record modeling. ALTA achieves superior performance in vision-language matching tasks like retrieval and zero-shot classification by adapting the pretrained vision model from masked record modeling. Additionally, we integrate temporal-multiview radiograph inputs to enhance the information consistency between radiographs and their corresponding descriptions in reports, further improving the vision-language alignment. Experimental evaluations show that ALTA outperforms the best-performing counterpart by over 4% absolute points in text-to-image accuracy and approximately 6% absolute points in image-to-text retrieval accuracy. The adaptation of vision-language models during efficient alignment also promotes better vision and language understanding. Code is publicly available at https://github.com/DopamineLcy/ALTA. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:02:27 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Lian",
"Chenyu",
""
],
[
"Zhou",
"Hong-Yu",
""
],
[
"Liang",
"Dongyun",
""
],
[
"Qin",
"Jing",
""
],
[
"Wang",
"Liansheng",
""
]
] |
2506.08991 | Anudeep Das | Anudeep Das, Gurjot Singh, Prach Chantasantitam, N. Asokan | Do Concept Replacement Techniques Really Erase Unacceptable Concepts? | null | null | null | null | cs.CV cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generative models, particularly diffusion-based text-to-image (T2I) models, have demonstrated astounding success. However, aligning them to avoid generating content with unacceptable concepts (e.g., offensive or copyrighted content, or celebrity likenesses) remains a significant challenge. Concept replacement techniques (CRTs) aim to address this challenge, often by trying to "erase" unacceptable concepts from models. Recently, model providers have started offering image editing services which accept an image and a text prompt as input, to produce an image altered as specified by the prompt. These are known as image-to-image (I2I) models. In this paper, we first use an I2I model to empirically demonstrate that today's state-of-the-art CRTs do not in fact erase unacceptable concepts. Existing CRTs are thus likely to be ineffective in emerging I2I scenarios, despite their proven ability to remove unwanted concepts in T2I pipelines, highlighting the need to understand this discrepancy between T2I and I2I settings. Next, we argue that a good CRT, while replacing unacceptable concepts, should preserve other concepts specified in the inputs to generative models. We call this fidelity. Prior work on CRTs have neglected fidelity in the case of unacceptable concepts. Finally, we propose the use of targeted image-editing techniques to achieve both effectiveness and fidelity. We present such a technique, AntiMirror, and demonstrate its viability. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:02:36 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Das",
"Anudeep",
""
],
[
"Singh",
"Gurjot",
""
],
[
"Chantasantitam",
"Prach",
""
],
[
"Asokan",
"N.",
""
]
] |
2506.08992 | Philippe Bergault | Philippe Bergault, Pierre Cardaliaguet, Wenbin Yan | Optimal hedging of an informed broker facing many traders | null | null | null | null | q-fin.TR math.OC | http://creativecommons.org/licenses/by/4.0/ | This paper investigates the optimal hedging strategies of an informed broker interacting with multiple traders in a financial market. We develop a theoretical framework in which the broker, possessing exclusive information about the drift of the asset's price, engages with traders whose trading activities impact the market price. Using a mean-field game approach, we derive the equilibrium strategies for both the broker and the traders, illustrating the intricate dynamics of their interactions. The broker's optimal strategy involves a Stackelberg equilibrium, where the broker leads and the traders follow. Our analysis also addresses the mean field limit of finite-player models and shows the convergence to the mean-field solution as the number of traders becomes large. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:04:39 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bergault",
"Philippe",
""
],
[
"Cardaliaguet",
"Pierre",
""
],
[
"Yan",
"Wenbin",
""
]
] |
2506.08993 | Jorge Terol Calvo | Toni Bert\'olez-Mart\'inez, Gonzalo Herrera, Pablo Mart\'inez-Mirav\'e, Jorge Terol Calvo | The Highest-Energy Neutrino Event Constrains Dark Matter-Neutrino Interactions | 16 pages, 5 figures | null | null | null | hep-ph astro-ph.CO astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Dark Matter-neutrino interactions affect the propagation of astrophysical neutrinos, attenuating the flux of neutrinos arriving at Earth. Using the highest-energy neutrino event detected to date by the KM3NeT collaboration as an example, and assuming an extragalactic origin, we derive limits on these interactions at $E_\nu = 220^{+570}_{-110}\, \mathrm{PeV}$. Considering only the propagation on the Milky Way Dark Matter halo, we constrain the interaction cross section over the mass of the Dark Matter candidate to be, $\sigma_{\rm DM-\nu}/m_{\rm DM} \lesssim 10^{-22}$~cm$^2$~GeV$^{-1}$. If a transient source was positively identified, the high-energy neutrino would have crossed the dark-matter halo of the source host as well, resulting in more stringent constraints (e.g., up to $ \sim$ 6 orders of magnitude assuming the blazar PKS 0605-085 is the source). These bounds on the Dark Matter-neutrino interaction cross section are translated into limits on the mass of the Dark Matter candidate, the mass of the mediator, and the coupling strength for different simplified models. In particular, we find that the KM3-230213A high-energy event sets the most stringent constraints on models where the scattering cross section rises in the high-energy limit. We identify vector Dark Matter interacting with neutrinos via vector mediators as the most promising scenario, and show that high-energy neutrinos provide world-leading limits in this model, testing the region of parameter space motivated by thermal freeze-out. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:11:22 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bertólez-Martínez",
"Toni",
""
],
[
"Herrera",
"Gonzalo",
""
],
[
"Martínez-Miravé",
"Pablo",
""
],
[
"Calvo",
"Jorge Terol",
""
]
] |
2506.08994 | Daniel Reichelt | Peter Meinzinger, Daniel Reichelt, and Federico Silvetti | Event generation at MEPS@NLO accuracy in neutral and charged current DIS at the EIC | 30 pages, 19 Figures | null | null | ZU-TH 43/25, CERN-TH-2025-110, IPPP/25/31, MCNET-25-13 | hep-ph hep-ex | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We present the first hadron-level predictions at MEPS@NLO precision for both neutral and charged current DIS at the EIC and the first application of the MEPS@NLO technique to charged current DIS in general. We critically examine state-of-the-art inclusive predictions using multi-leg merging techniques, contrasting perturbative and non-perturbative uncertainties. We study typical kinematic DIS observables as well as jet measurements and 1-jettiness with realistic cuts implied by expected and past detector resolution. On the perturbative side, we see large corrections towards small virtualities and Bjorken-x, which can be captured by higher-multiplicity matrix elements and the merging procedure. Non-perturbative effects, while negligible in most jet observables, can reach similar size as the perturbative uncertainties around the peak of the 1-jettiness distributions especially at low values of $Q^2$. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:13:15 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Meinzinger",
"Peter",
""
],
[
"Reichelt",
"Daniel",
""
],
[
"Silvetti",
"Federico",
""
]
] |
2506.08995 | Dipanjan Mitra | Dipanjan Mitra, Mattia Negrello, Gianfranco De Zotti | Investigating the physical properties of dusty star-forming galaxies at z>=1.5 in the GOODS-South field using JWST | 23 pages, 14 figures, accepted for publication in ApJ | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We investigated how well the physical properties of progenitors of present-day massive spheroidal galaxies (proto-spheroids) can be constrained by the JWST Advanced Deep Extragalactic Survey (JADES) in the GOODS-South field, which benefits from extensive photometric and spectroscopic data, including those from the Hubble, Spitzer and Herschel. We adopted a physical model for the evolution of proto-spheroidal galaxies, which form the bulk of dusty star-forming galaxies (DSFGs) at z>=1.5 and confirmed its consistency with recent mid-infrared high-z galaxy luminosity functions. Using the model and the JADES survey strategy, we simulated a sample of proto-spheroids over 87.5 arcmin^2, matching the JADES/GOODS-S survey area. Photometric redshifts estimated from simulated JWST photometry showed >=95% accuracy and were used in SED fitting with CIGALE. We demonstrated that JWST will provide reliable stellar mass estimates up to 0.1 dex for the majority of proto-spheroids at z>=1.5 and can detect low-mass systems during cosmic noon that were inaccessible in the pre-JWST era. Focusing on the active star-forming phase of the proto-spheroid evolution, we defined a sub-sample flux limited at 250 micron (DSFG sample) and derived SFR, dust luminosity and dust mass complementing the JWST photometry with that from Spitzer/MIPS and Herschel. We also constructed a JWST-selected DSFG catalog from ASTRODEEP data using NIRCam colour criteria and demonstrated strong consistency between the observed and simulated DSFG populations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:13:21 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Mitra",
"Dipanjan",
""
],
[
"Negrello",
"Mattia",
""
],
[
"De Zotti",
"Gianfranco",
""
]
] |
2506.08996 | Brian Tang | Brian Tang, Duc Bui, Kang G. Shin | Navigating Cookie Consent Violations Across the Globe | Published at 34th USENIX Security Symposium (2025) | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Online services provide users with cookie banners to accept/reject the cookies placed on their web browsers. Despite the increased adoption of cookie banners, little has been done to ensure that cookie consent is compliant with privacy laws around the globe. Prior studies have found that cookies are often placed on browsers even after their explicit rejection by users. These inconsistencies in cookie banner behavior circumvent users' consent preferences and are known as cookie consent violations. To address this important problem, we propose an end-to-end system, called ConsentChk, that detects and analyzes cookie banner behavior. ConsentChk uses a formal model to systematically detect and categorize cookie consent violations. We investigate eight English-speaking regions across the world, and analyze cookie banner behavior across 1,793 globally-popular websites. Cookie behavior, cookie consent violation rates, and cookie banner implementations are found to be highly dependent on region. Our evaluation reveals that consent management platforms (CMPs) and website developers likely tailor cookie banner configurations based on their (often incorrect) interpretations of regional privacy laws. We discuss various root causes behind these cookie consent violations. The resulting implementations produce misleading cookie banners, indicating the prevalence of inconsistently implemented and enforced cookie consent between various regions. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:15:12 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Tang",
"Brian",
""
],
[
"Bui",
"Duc",
""
],
[
"Shin",
"Kang G.",
""
]
] |
2506.08997 | Fabian Immel | Fabian Immel, Jan-Hendrik Pauls, Richard Fehler, Frank Bieder, Jonas Merkert, Christoph Stiller | SDTagNet: Leveraging Text-Annotated Navigation Maps for Online HD Map Construction | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Autonomous vehicles rely on detailed and accurate environmental information to operate safely. High definition (HD) maps offer a promising solution, but their high maintenance cost poses a significant barrier to scalable deployment. This challenge is addressed by online HD map construction methods, which generate local HD maps from live sensor data. However, these methods are inherently limited by the short perception range of onboard sensors. To overcome this limitation and improve general performance, recent approaches have explored the use of standard definition (SD) maps as prior, which are significantly easier to maintain. We propose SDTagNet, the first online HD map construction method that fully utilizes the information of widely available SD maps, like OpenStreetMap, to enhance far range detection accuracy. Our approach introduces two key innovations. First, in contrast to previous work, we incorporate not only polyline SD map data with manually selected classes, but additional semantic information in the form of textual annotations. In this way, we enrich SD vector map tokens with NLP-derived features, eliminating the dependency on predefined specifications or exhaustive class taxonomies. Second, we introduce a point-level SD map encoder together with orthogonal element identifiers to uniformly integrate all types of map elements. Experiments on Argoverse 2 and nuScenes show that this boosts map perception performance by up to +5.9 mAP (+45%) w.r.t. map construction without priors and up to +3.2 mAP (+20%) w.r.t. previous approaches that already use SD map priors. Code is available at https://github.com/immel-f/SDTagNet | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:16:00 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Immel",
"Fabian",
""
],
[
"Pauls",
"Jan-Hendrik",
""
],
[
"Fehler",
"Richard",
""
],
[
"Bieder",
"Frank",
""
],
[
"Merkert",
"Jonas",
""
],
[
"Stiller",
"Christoph",
""
]
] |
2506.08998 | Gilles Bareilles | Gilles Bareilles, Julien Fageot, L\^e-Nguy\^en Hoang, Peva Blanchard, Wassim Bouaziz, S\'ebastien Rouault, El-Mahdi El-Mhamdi | On Monotonicity in AI Alignment | null | null | null | null | math.ST stat.TH | http://creativecommons.org/licenses/by/4.0/ | Comparison-based preference learning has become central to the alignment of AI models with human preferences. However, these methods may behave counterintuitively. After empirically observing that, when accounting for a preference for response $y$ over $z$, the model may actually decrease the probability (and reward) of generating $y$ (an observation also made by others), this paper investigates the root causes of (non) monotonicity, for a general comparison-based preference learning framework that subsumes Direct Preference Optimization (DPO), Generalized Preference Optimization (GPO) and Generalized Bradley-Terry (GBT). Under mild assumptions, we prove that such methods still satisfy what we call local pairwise monotonicity. We also provide a bouquet of formalizations of monotonicity, and identify sufficient conditions for their guarantee, thereby providing a toolbox to evaluate how prone learning models are to monotonicity violations. These results clarify the limitations of current methods and provide guidance for developing more trustworthy preference learning algorithms. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:17:48 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bareilles",
"Gilles",
""
],
[
"Fageot",
"Julien",
""
],
[
"Hoang",
"Lê-Nguyên",
""
],
[
"Blanchard",
"Peva",
""
],
[
"Bouaziz",
"Wassim",
""
],
[
"Rouault",
"Sébastien",
""
],
[
"El-Mhamdi",
"El-Mahdi",
""
]
] |
2506.08999 | Theo Zhang | Theo Zhang, Madurya Suresh, Anne S. Warlaumont, Kasia Hitczenko, Alejandrina Cristia, Margaret Cychosz | Employing self-supervised learning models for cross-linguistic child speech maturity classification | To be published in Interspeech 2025. 5 pages, 2 figures. For associated Github repository, see https://github.com/spoglab-stanford/w2v2-pro-sm/tree/main/speechbrain/recipes/W2V2-LL4300-Pro-SM | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Speech technology systems struggle with many downstream tasks for child speech due to small training corpora and the difficulties that child speech pose. We apply a novel dataset, SpeechMaturity, to state-of-the-art transformer models to address a fundamental classification task: identifying child vocalizations. Unlike previous corpora, our dataset captures maximally ecologically-valid child vocalizations across an unprecedented sample, comprising children acquiring 25+ languages in the U.S., Bolivia, Vanuatu, Papua New Guinea, Solomon Islands, and France. The dataset contains 242,004 labeled vocalizations, magnitudes larger than previous work. Models were trained to distinguish between cry, laughter, mature (consonant+vowel), and immature speech (just consonant or vowel). Models trained on the dataset outperform state-of-the-art models trained on previous datasets, achieved classification accuracy comparable to humans, and were robust across rural and urban settings. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:20:02 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhang",
"Theo",
""
],
[
"Suresh",
"Madurya",
""
],
[
"Warlaumont",
"Anne S.",
""
],
[
"Hitczenko",
"Kasia",
""
],
[
"Cristia",
"Alejandrina",
""
],
[
"Cychosz",
"Margaret",
""
]
] |
2506.09000 | David Jekel | Ian Charlesworth, David Jekel | The atoms of graph product von Neumann algebras | 22 pages | null | null | null | math.OA math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We completely classify the atomic summands in a graph product $(M,\varphi) = *_{v \in \mathcal{G}} (M_v,\varphi_v)$ of von Neumann algebras with faithful normal states. Each type I factor summand $(N,\psi)$ is a tensor product of type I factor summands $(N_v,\psi_v)$ in the individual algebras. The existence of such a summand and its weight in the direct sum can be determined from the $(N_v,\psi_v)$'s using explicit polynomials associated to the graph. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:20:17 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Charlesworth",
"Ian",
""
],
[
"Jekel",
"David",
""
]
] |
2506.09001 | Jonathan Franceschi | Jonathan Franceschi and Lorenzo Pareschi and Mattia Zanella | Superlinear Drift in Consensus-Based Optimization with Condensation Phenomena | null | null | null | null | math.OC math.AP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Consensus-based optimization (CBO) is a class of metaheuristic algorithms designed for global optimization problems. In the many-particle limit, classical CBO dynamics can be rigorously connected to mean-field equations that ensure convergence toward global minimizers under suitable conditions. In this work, we draw inspiration from recent extensions of the Kaniadakis--Quarati model for indistinguishable bosons to develop a novel CBO method governed by a system of SDEs with superlinear drift and nonconstant diffusion. The resulting mean-field formulation in one dimension exhibits condensation-like phenomena, including finite-time blow-up and loss of $L^2$-regularity. To avoid the curse of dimensionality a marginal based formulation which permits to leverage the one-dimensional results to multiple dimensions is proposed. We support our approach with numerical experiments that highlight both its consistency and potential performance improvements compared to classical CBO methods. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:21:20 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Franceschi",
"Jonathan",
""
],
[
"Pareschi",
"Lorenzo",
""
],
[
"Zanella",
"Mattia",
""
]
] |
2506.09004 | Bengt J. Nilsson | Andrej Brodnik and Bengt J. Nilsson and Gordana Vujovi\'c | Improving Online Bin Covering with Little Advice | 13 pages. This is an extended version of an article published in "35th International Workshop on Combinatorial Algorithms. IWOCA 2024." Lecture Notes in Computer Science, vol 14764. Springer Verlag | null | 10.1007/978-3-031-63021-7_6 | null | cs.DS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The online bin covering problem is: given an input sequence of items find a placement of the items in the maximum number of bins such that the sum of the items' sizes in each bin is at least~1. Boyar~{\em et~al}.\@~\cite{boyar2021} present a strategy that with $O(\log \log n)$ bits of advice, where $n$ is the length of the input sequence, achieves a competitive ratio of $8/15\approx0.5333\ldots$. We show that with a strengthened analysis and some minor improvements, the same strategy achieves the significantly improved competitive ratio of~$135/242\approx0.5578\ldots$, still using $O(\log \log n)$ bits of advice. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:23:46 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Brodnik",
"Andrej",
""
],
[
"Nilsson",
"Bengt J.",
""
],
[
"Vujović",
"Gordana",
""
]
] |
2506.09005 | Alexis Michelat | Alexis Michelat | Morse Index Stability of Branched Willmore Immersions | 131 pages | null | null | null | math.DG math.AP | http://creativecommons.org/licenses/by/4.0/ | We show the Morse index stability for sequences of Willmore immersions of bounded energy without assuming that the limiting immersion and the bubbles are free of branch points. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:26:28 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Michelat",
"Alexis",
""
]
] |
2506.09006 | Harri Waltari | Stefano Moretti, Luca Panizzi, J\"orgen Sj\"olin and Harri Waltari | Deconstructing resonant Higgs pair production at the LHC: effects of coloured and neutral scalars in the NMSSM test case | 20 pages, 9 figures, UFO model used in this work is available upon request and will be made public soon | null | null | null | hep-ph hep-ex | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We study resonant production of pairs of Standard Model (SM)-like Higgs bosons, in the presence of new neutral Higgs states together with new coloured scalars (stops or sbottoms) in loops within the Next-to-Minimal Supersymmetric SM (NMSSM). This is used as a test case to prove that the Large Hadron Collider has sensitivity to a variety of effects stemming from interferences between resonant (heavy) Higgs diagrams and/or among these and non-resonant topologies involving loops of both tops and stops. These effects can alter significantly the naive description of individual $s$-channel Breit-Wigner resonances, leading to distortions of the latter which, on the one hand, may mask their presence but, on the other hand, could enable one to extract features of the underlying new physics scenario. This last aspect is made possible through a decomposition of the $gg\to hh$ signal process into all its amplitude components, each of which has a well-defined coupling structure. Ultimately, such effects can be traced back to the relevant Feynman diagrams and can enable a detailed interpretation of this process. To illustrate this, we introduce various Benchmark Points that exhibit potentially observable features during the current and/or upcoming runs of the LHC in one or more of the three customary di-Higgs decay channels: $b\bar bb\bar b$, $b\bar b \tau^+\tau^-$ and $b\bar b\gamma\gamma$. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:27:49 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Moretti",
"Stefano",
""
],
[
"Panizzi",
"Luca",
""
],
[
"Sjölin",
"Jörgen",
""
],
[
"Waltari",
"Harri",
""
]
] |
2506.09008 | Eugene DePrince | Samragni Banerjee and Run R. Li and Brandon C. Cooper and Tianyuan Zhang and Edward F. Valeev and Xiaosong Li and A. Eugene DePrince III | Relativistic Core-Valence-Separated Molecular Mean-Field Exact-Two-Component Equation-of-Motion Coupled Cluster Theory: Applications to L-edge X-ray Absorption Spectroscopy | null | null | null | null | physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | L-edge X-ray absorption spectra for first-row transition metal complexes are obtained from relativistic equation-of-motion singles and doubles coupled-cluster (EOM-CCSD) calculations that make use of the core-valence separation (CVS) scheme, with scalar and spin--orbit relativistic effects modeled within the molecular mean-field exact two-component (X2C) framework. By incorporating relativistic effects variationally at the Dirac--Coulomb--Breit (DCB) reference level, this method delivers accurate predictions of L-edge features, including energy shifts, intensity ratios, and fine-structure splittings, across a range of molecular systems. Benchmarking against perturbative spin--orbit treatments and relativistic TDDFT highlights the superior performance and robustness of the CVS-DCB-X2C-EOM-CCSD approach, including the reliability of basis set recontraction schemes. While limitations remain in describing high-density spectral regions, our results establish CVS-DCB-X2C-EOM-CCSD as a powerful and broadly applicable tool for relativistic core-excitation spectroscopy. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:34:56 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Banerjee",
"Samragni",
""
],
[
"Li",
"Run R.",
""
],
[
"Cooper",
"Brandon C.",
""
],
[
"Zhang",
"Tianyuan",
""
],
[
"Valeev",
"Edward F.",
""
],
[
"Li",
"Xiaosong",
""
],
[
"DePrince",
"A. Eugene",
"III"
]
] |
2506.09010 | Sebastian Schmidt | Sebastian Schmidt and Prasanga Dhungel and Christoffer L\"offler and Bj\"orn Nieth and Stephan G\"unnemann and Leo Schwinn | Effective Data Pruning through Score Extrapolation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training advanced machine learning models demands massive datasets, resulting in prohibitive computational costs. To address this challenge, data pruning techniques identify and remove redundant training samples while preserving model performance. Yet, existing pruning techniques predominantly require a full initial training pass to identify removable samples, negating any efficiency benefits for single training runs. To overcome this limitation, we introduce a novel importance score extrapolation framework that requires training on only a small subset of data. We present two initial approaches in this framework - k-nearest neighbors and graph neural networks - to accurately predict sample importance for the entire dataset using patterns learned from this minimal subset. We demonstrate the effectiveness of our approach for 2 state-of-the-art pruning methods (Dynamic Uncertainty and TDDS), 4 different datasets (CIFAR-10, CIFAR-100, Places-365, and ImageNet), and 3 training paradigms (supervised, unsupervised, and adversarial). Our results indicate that score extrapolation is a promising direction to scale expensive score calculation methods, such as pruning, data attribution, or other tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:38:49 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Schmidt",
"Sebastian",
""
],
[
"Dhungel",
"Prasanga",
""
],
[
"Löffler",
"Christoffer",
""
],
[
"Nieth",
"Björn",
""
],
[
"Günnemann",
"Stephan",
""
],
[
"Schwinn",
"Leo",
""
]
] |
2506.09011 | Jiaozi Wang | Jiaozi Wang, Hua Yan, Robin Steinigeweg, Jochen Gemmer | Eigenstate Thermalization Hypothesis and Random Matrix Theory Universality in Few-Body Systems | 13 pages, 14 figures | null | null | null | cond-mat.stat-mech nlin.CD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we study the Feingold-Peres model as an example, which is a well-known paradigm of quantum chaos. Using semiclassical analysis and numerical simulations, we study the statistical properties of observables in few-body systems with chaotic classical limits and the emergence of random matrix theory universality. More specifically, we focus on: 1) the applicability of the eigenstate thermalization hypothesis in few-body systems and the dependence of its form on the effective Planck constant and 2) the existence of a universal random matrix theory description of observables when truncated to a small microcanonical energy window. Our results provide new insights into the established field of few-body quantum chaos and help bridge it to modern perspectives, such as the general eigenstate thermalization hypothesis (ETH). | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:41:08 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Jiaozi",
""
],
[
"Yan",
"Hua",
""
],
[
"Steinigeweg",
"Robin",
""
],
[
"Gemmer",
"Jochen",
""
]
] |
2506.09013 | Idrees Qasim Dr | Idrees Qasim | Generalization of Cauchy type theorems for matrix Polynomials | null | null | null | null | math.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, we find bounds for the eigenvalues of matrix polynomials. In particular, we find generalizations of Cauchy's classical Theorem for distribution of eigenvalues of matrix polynomial. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:42:06 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Qasim",
"Idrees",
""
]
] |
2506.09014 | Jianing Qi | Jianing Qi, Xi Ye, Hao Tang, Zhigang Zhu, Eunsol Choi | Learning to Reason Across Parallel Samples for LLM Reasoning | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Scaling test-time compute brings substantial performance gains for large language models (LLMs). By sampling multiple answers and heuristically aggregate their answers (e.g., either through majority voting or using verifiers to rank the answers), one can achieve consistent performance gains in math domains. In this paper, we propose a new way to leverage such multiple sample set. We train a compact LLM, called Sample Set Aggregator (SSA), that takes a concatenated sequence of multiple samples and output the final answer, optimizing it for the answer accuracy with reinforcement learning. Experiments on multiple reasoning datasets show that SSA outperforms other test-time scaling methods such as reward model-based re-ranking. Our approach also shows a promising generalization ability, across sample set sizes, base model families and scales, and tasks. By separating LLMs to generate answers and LLMs to analyze and aggregate sampled answers, our approach can work with the outputs from premier black box models easily and efficiently. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:42:35 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Qi",
"Jianing",
""
],
[
"Ye",
"Xi",
""
],
[
"Tang",
"Hao",
""
],
[
"Zhu",
"Zhigang",
""
],
[
"Choi",
"Eunsol",
""
]
] |
2506.09015 | George H. Seelinger | Jonah Blasiak, Mark Haiman, Jennifer Morse, Anna Pun, George H. Seelinger | Flagged LLT polynomials, nonsymmetric plethysm, and nonsymmetric Macdonald polynomials | null | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | The plethystic transformation $f[X] \mapsto f[X/(1-t)]$ and LLT polynomials are central to the theory of symmetric Macdonald polynomials. In this work, we introduce and study nonsymmetric flagged LLT polynomials. We show that these admit both an algebraic and a combinatorial description, that they Weyl symmetrize to the usual symmetric LLT polynomials, and we conjecture that they expand positively in terms of Demazure atoms. Additionally, we construct a nonsymmetric plethysm operator $\Pi_{t,x}$ on $\mathfrak{K}[x_1,\ldots,x_n]$, which serves as an analogue of $f[X] \mapsto f[X/(1-t)]$. We prove that $\Pi_{t,x}$ remarkably maps flagged LLT polynomials defined over a signed alphabet to ones over an unsigned alphabet.
Our main application of this theory is to formulate a nonsymmetric version of Macdonald positivity, similar in spirit to conjectures of Knop and Lapointe, but with several new features. To do this, we recast the Haglund-Haiman-Loehr formula for nonsymmetric Macdonald polynomials $\mathcal{E}_{\mu }(x;q,t)$ as a positive sum of signed flagged LLT polynomials. Then, after taking a suitable stable limit of both $\mathcal{E}_{\mu }(x;q,t)$ and $\Pi_{t,x}$, we obtain modified nonsymmetric Macdonald polynomials which are positive sums of flagged LLT polynomials and thus are conjecturally atom positive, strengthening the Macdonald positivity conjecture. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:42:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Blasiak",
"Jonah",
""
],
[
"Haiman",
"Mark",
""
],
[
"Morse",
"Jennifer",
""
],
[
"Pun",
"Anna",
""
],
[
"Seelinger",
"George H.",
""
]
] |
2506.09016 | Daman Arora | Ruiqi Zhang, Daman Arora, Song Mei, Andrea Zanette | SPEED-RL: Faster Training of Reasoning Models via Online Curriculum Learning | pre-print | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Training large language models with reinforcement learning (RL) against verifiable rewards significantly enhances their reasoning abilities, yet remains computationally expensive due to inefficient uniform prompt sampling. We introduce Selective Prompting with Efficient Estimation of Difficulty (SPEED), an adaptive online RL curriculum that selectively chooses training examples of intermediate difficulty to maximize learning efficiency. Theoretically, we establish that intermediate-difficulty prompts improve the gradient estimator's signal-to-noise ratio, accelerating convergence. Empirically, our efficient implementation leads to 2x to 6x faster training without degrading accuracy, requires no manual tuning, and integrates seamlessly into standard RL algorithms. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:42:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhang",
"Ruiqi",
""
],
[
"Arora",
"Daman",
""
],
[
"Mei",
"Song",
""
],
[
"Zanette",
"Andrea",
""
]
] |
2506.09017 | Hiram H. L\'opez | Daniel P. Bossaller and Hiram H. L\'opez | Linear exact repair schemes for free MDS and Reed-Solomon codes over Galois rings | null | null | null | null | cs.IT math.IT math.RA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Codes over rings, especially over Galois rings, have been extensively studied for nearly three decades due to their similarity to linear codes over finite fields. A distributed storage system uses a linear code to encode a large file across several nodes. If one of the nodes fails, a linear exact repair scheme efficiently recovers the failed node by accessing and downloading data from the rest of the servers of the storage system. In this article, we develop a linear repair scheme for free maximum distance separable codes, which coincide with free maximum distance with respect to the rank codes over Galois rings. In particular, we give a linear repair scheme for full-length Reed-Solomon codes over a Galois ring. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:43:04 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Bossaller",
"Daniel P.",
""
],
[
"López",
"Hiram H.",
""
]
] |
2506.09018 | Marton Havasi | Marton Havasi, Brian Karrer, Itai Gat, Ricky T. Q. Chen | Edit Flows: Flow Matching with Edit Operations | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Autoregressive generative models naturally generate variable-length sequences, while non-autoregressive models struggle, often imposing rigid, token-wise structures. We propose Edit Flows, a non-autoregressive model that overcomes these limitations by defining a discrete flow over sequences through edit operations-insertions, deletions, and substitutions. By modeling these operations within a Continuous-time Markov Chain over the sequence space, Edit Flows enable flexible, position-relative generation that aligns more closely with the structure of sequence data. Our training method leverages an expanded state space with auxiliary variables, making the learning process efficient and tractable. Empirical results show that Edit Flows outperforms both autoregressive and mask models on image captioning and significantly outperforms the mask construction in text and code generation. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:44:19 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Havasi",
"Marton",
""
],
[
"Karrer",
"Brian",
""
],
[
"Gat",
"Itai",
""
],
[
"Chen",
"Ricky T. Q.",
""
]
] |
2506.09020 | Zichao Dong | Zichao Dong, Jun Gao, Ruonan Li, Hong Liu | Induced rational exponents and bipartite subgraphs in $K_{s, s}$-free graphs | 21 pages, 3 figures | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | In this paper, we study a general phenomenon that many extremal results for bipartite graphs can be transferred to the induced setting when the host graph is $K_{s, s}$-free. As manifestations of this phenomenon, we prove that every rational $\frac{a}{b} \in (1, 2), \, a, b \in \mathbb{N}_+$, can be achieved as Tur\'{a}n exponent of a family of at most $2^a$ induced forbidden bipartite graphs, extending a result of Bukh and Conlon [JEMS 2018]. Our forbidden family is a subfamily of theirs which is substantially smaller. A key ingredient, which is yet another instance of this phenomenon, is supersaturation results for induced trees and cycles in $K_{s, s}$-free graphs. We also provide new evidence to a recent conjecture of Hunter, Milojevi\'{c}, Sudakov, and Tomon [JCTB 2025] by proving optimal bounds for the maximum size of $K_{s, s}$-free graphs without an induced copy of theta graphs or prism graphs, whose Tur\'{a}n exponents were determined by Conlon [BLMS 2019] and by Gao, Janzer, Liu, and Xu [IJM 2025+]. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:45:47 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Dong",
"Zichao",
""
],
[
"Gao",
"Jun",
""
],
[
"Li",
"Ruonan",
""
],
[
"Liu",
"Hong",
""
]
] |
2506.09024 | Felix Wagner MSc. | Felix Wagner, Pramit Saha, Harry Anthony, J. Alison Noble, Konstantinos Kamnitsas | DIsoN: Decentralized Isolation Networks for Out-of-Distribution Detection in Medical Imaging | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Safe deployment of machine learning (ML) models in safety-critical domains such as medical imaging requires detecting inputs with characteristics not seen during training, known as out-of-distribution (OOD) detection, to prevent unreliable predictions. Effective OOD detection after deployment could benefit from access to the training data, enabling direct comparison between test samples and the training data distribution to identify differences. State-of-the-art OOD detection methods, however, either discard training data after deployment or assume that test samples and training data are centrally stored together, an assumption that rarely holds in real-world settings. This is because shipping training data with the deployed model is usually impossible due to the size of training databases, as well as proprietary or privacy constraints. We introduce the Isolation Network, an OOD detection framework that quantifies the difficulty of separating a target test sample from the training data by solving a binary classification task. We then propose Decentralized Isolation Networks (DIsoN), which enables the comparison of training and test data when data-sharing is impossible, by exchanging only model parameters between the remote computational nodes of training and deployment. We further extend DIsoN with class-conditioning, comparing a target sample solely with training data of its predicted class. We evaluate DIsoN on four medical imaging datasets (dermatology, chest X-ray, breast ultrasound, histopathology) across 12 OOD detection tasks. DIsoN performs favorably against existing methods while respecting data-privacy. This decentralized OOD detection framework opens the way for a new type of service that ML developers could provide along with their models: providing remote, secure utilization of their training data for OOD detection services. Code will be available upon acceptance at: ***** | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:52:18 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wagner",
"Felix",
""
],
[
"Saha",
"Pramit",
""
],
[
"Anthony",
"Harry",
""
],
[
"Noble",
"J. Alison",
""
],
[
"Kamnitsas",
"Konstantinos",
""
]
] |
2506.09025 | Yiping Ma | Yi-Ping Ma, Ivan Sudakow, P. L. Krapivsky | Mixed phases in feedback Ising models | 7 pages, 3 figures | null | null | null | cond-mat.stat-mech math.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study mean-field Ising models whose coupling depends on the magnetization via a feedback function. We identify mixed phases (MPs) and show that they can be stable at zero temperature for sufficiently strong feedback. Moreover, stable MPs are always super-stable with perturbation decaying linearly in time. We argue that such feedback Ising models (FIMs) provide a useful framework for phase transformations between aligned phases via stable and unstable intermediate phases in multistable systems. We also analyze the dynamical behavior of FIMs driven by a time-varying magnetic field. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:52:28 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Ma",
"Yi-Ping",
""
],
[
"Sudakow",
"Ivan",
""
],
[
"Krapivsky",
"P. L.",
""
]
] |
2506.09026 | Amrith Setlur | Amrith Setlur, Matthew Y. R. Yang, Charlie Snell, Jeremy Greer, Ian Wu, Virginia Smith, Max Simchowitz, Aviral Kumar | e3: Learning to Explore Enables Extrapolation of Test-Time Compute for LLMs | null | null | null | null | cs.LG cs.CL | http://creativecommons.org/licenses/by/4.0/ | Test-time scaling offers a promising path to improve LLM reasoning by utilizing more compute at inference time; however, the true promise of this paradigm lies in extrapolation (i.e., improvement in performance on hard problems as LLMs keep "thinking" for longer, beyond the maximum token budget they were trained on). Surprisingly, we find that most existing reasoning models do not extrapolate well. We show that one way to enable extrapolation is by training the LLM to perform in-context exploration: training the LLM to effectively spend its test time budget by chaining operations (such as generation, verification, refinement, etc.), or testing multiple hypotheses before it commits to an answer. To enable in-context exploration, we identify three key ingredients as part of our recipe e3: (1) chaining skills that the base LLM has asymmetric competence in, e.g., chaining verification (easy) with generation (hard), as a way to implement in-context search; (2) leveraging "negative" gradients from incorrect traces to amplify exploration during RL, resulting in longer search traces that chains additional asymmetries; and (3) coupling task difficulty with training token budget during training via a specifically-designed curriculum to structure in-context exploration. Our recipe e3 produces the best known 1.7B model according to AIME'25 and HMMT'25 scores, and extrapolates to 2x the training token budget. Our e3-1.7B model not only attains high pass@1 scores, but also improves pass@k over the base model. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:52:42 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Setlur",
"Amrith",
""
],
[
"Yang",
"Matthew Y. R.",
""
],
[
"Snell",
"Charlie",
""
],
[
"Greer",
"Jeremy",
""
],
[
"Wu",
"Ian",
""
],
[
"Smith",
"Virginia",
""
],
[
"Simchowitz",
"Max",
""
],
[
"Kumar",
"Aviral",
""
]
] |
2506.09027 | Runqian Wang | Runqian Wang, Kaiming He | Diffuse and Disperse: Image Generation with Representation Regularization | null | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | The development of diffusion-based generative models over the past decade has largely proceeded independently of progress in representation learning. These diffusion models typically rely on regression-based objectives and generally lack explicit regularization. In this work, we propose \textit{Dispersive Loss}, a simple plug-and-play regularizer that effectively improves diffusion-based generative models. Our loss function encourages internal representations to disperse in the hidden space, analogous to contrastive self-supervised learning, with the key distinction that it requires no positive sample pairs and therefore does not interfere with the sampling process used for regression. Compared to the recent method of representation alignment (REPA), our approach is self-contained and minimalist, requiring no pre-training, no additional parameters, and no external data. We evaluate Dispersive Loss on the ImageNet dataset across a range of models and report consistent improvements over widely used and strong baselines. We hope our work will help bridge the gap between generative modeling and representation learning. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:53:29 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Runqian",
""
],
[
"He",
"Kaiming",
""
]
] |
2506.09028 | Stephan Tasler | Stephan Tasler, Josias Old, Lukas Heunisch, Verena Feulner, Timo Eckstein, Markus M\"uller and Michael J. Hartmann | Optimizing Superconducting Three-Qubit Gates for Surface-Code Error Correction | 12 pages, 11 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Quantum error correction (QEC) is one of the crucial building blocks for developing quantum computers that have significant potential for reaching a quantum advantage in applications. Prominent candidates for QEC are stabilizer codes for which periodic readout of stabilizer operators is typically implemented via successive two-qubit entangling gates, and is repeated many times during a computation. To improve QEC performance, it is thus beneficial to make the stabilizer readout faster and less prone to fault-tolerance-breaking errors. Here we design a 3-qubit CZZ gate for superconducting transmon qubits that maps the parity of two data qubits onto one measurement qubit in a single step. We find that the gate can be executed in a duration of $35\,$ns with a fidelity of F$=99.96 \, \%$. To optimize the gate, we use an error model obtained from the microscopic gate simulation to systematically suppress Pauli errors that are particularly harmful to the QEC protocol. Using this error model, we investigate the implementation of this 3-qubit gate in a surface code syndrome readout schedule. We find that for the rotated surface code, the implementation of CZZ gates increases the error threshold by nearly 50\% to $\approx 1.2\,\%$ and decreases the logical error rate, in the experimental relevant regime, by up to one order of magnitude, compared to the standard CZ readout protocol. We also show that for the unrotated surface code, strictly fault-tolerant readout schedules can be found. This opens a new perspective for below-threshold surface-code error correction, where it can be advantageous to use multi-qubit gates instead of two-qubit gates to obtain a better QEC performance. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:54:22 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Tasler",
"Stephan",
""
],
[
"Old",
"Josias",
""
],
[
"Heunisch",
"Lukas",
""
],
[
"Feulner",
"Verena",
""
],
[
"Eckstein",
"Timo",
""
],
[
"Müller",
"Markus",
""
],
[
"Hartmann",
"Michael J.",
""
]
] |
2506.09029 | Josias Old | Josias Old, Stephan Tasler, Michael J. Hartmann, Markus M\"uller | Fault-Tolerant Stabilizer Measurements in Surface Codes with Three-Qubit Gates | 7 pages, 6 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Quantum error correction (QEC) is considered a deciding component in enabling practical quantum computing. Stabilizer codes, and in particular topological surface codes, are promising candidates for implementing QEC by redundantly encoding quantum information. While it is widely believed that a strictly fault-tolerant protocol can only be implemented using single- and two-qubit gates, several quantum computing platforms, based on trapped ions, neutral atoms and also superconducting qubits support native multi-qubit operations, e.g. using multi-ion entangling gates, Rydberg blockade or parallelized tunable couplers, respectively. In this work, we show that stabilizer measurement circuits for unrotated surface codes can be fault-tolerant using single auxiliary qubits and three-qubit gates. These gates enable lower-depth circuits leading to fewer fault locations and potentially shorter QEC cycle times. Concretely, we find that in an optimistic parameter regime where fidelities of three-qubit gates are the same as those of two-qubit gates, the logical error rate can be up to one order of magnitude lower and the threshold can be significantly higher, increasing from $\approx 0.76 \%$ to $\approx 1.05 \%$. Our results, which are applicable to a wide range of platforms, thereby motivate further investigation into multi-qubit gates as components for fault-tolerant QEC, as they can lead to substantial advantages in terms of time and physical qubit resources required to reach a target logical error rate. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:54:23 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Old",
"Josias",
""
],
[
"Tasler",
"Stephan",
""
],
[
"Hartmann",
"Michael J.",
""
],
[
"Müller",
"Markus",
""
]
] |
2506.09030 | Adham Naji | Adham Naji, Pawan Kumar Gupta, and Gennady Stupakov | Full-scatter vector field analysis of an overmoded and periodically-loaded cylindrical structure for the transportation of THz radiation | 17 pages, 13 figures | null | null | null | physics.acc-ph math-ph math.MP physics.app-ph physics.class-ph physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The full-scatter vector field analysis of a highly overmoded and periodically loaded cylindrical wavguiding structure (or "iris line") is presented. This type of structure is promising for the transportation of diffraction-prone short THz radiation pulses over hundreds of meters. The analysis uses vector fields, superseding scalar field descriptions, to account for polarization effects and ohmic loss on practical conductive screens. Its results are quite general, as it only assumes a paraxial incidence with slowly-varying envelope (parabolic wave equation) along the axis of the structure. It removes the traditional assumption of very thin screens, allowing for the study of thicker screens in the high-frequency limit, while hosting hundreds of modes and formulating the problem efficiently by scattering matrices whose coefficients are found analytically. The theory expands previously established methods, including Vainstein's limit and the forward-scatter approximation, and validates them (as they become special cases with respect to it). It also facilitates accurate visualization of the transient regime at the entrance of the structure, and how it evolves to reach steady state. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:54:33 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Naji",
"Adham",
""
],
[
"Gupta",
"Pawan Kumar",
""
],
[
"Stupakov",
"Gennady",
""
]
] |
2506.09032 | Jonatan Herrera | J\'onatan Herrera and Miguel S\'anchez | Finslerian lightconvex boundaries: applications to causal simplicity and the space of cone geodesics $\mathcal{N}$ | 37 pages, 5 figures | null | null | null | math.DG gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Our outcome is structured in the following sequence: (1) a general result for indefinite Finslerian manifolds with boundary $(M,L)$ showing the equivalence between local and infinitesimal (time, light or space) convexities for the boundary $\partial M$, (2) for any cone structure $(M,\mathcal{C})$ which is globally hyperbolic with timelike boundary, the equivalence among: (a) the boundary $\partial M$ is lightconvex, (b) the interior $\mathring{M}$ is causally simple and (c) the space of the cone (null) geodesics $\mathcal{N}$ of $(\mathring{M},\mathcal{C})$ is Hausdorff, (3) in this case, the manifold structure of $\mathcal{N}$ is obtained explicitly in terms of elements in $\partial M$ and a smooth Cauchy hypersurface $S$, (4) the known results and examples about Hausdorfness of $\mathcal{N}$ are revisited and extended, leading to the notion of {\em causally simple spacetime with $T_2$-lightspace} as a step in the causal ladder below global hyperbolicity.
The results are significant for relativistic (Lorentz) spacetimes and the writing allows one either to be introduced in Finslerian technicalities or to skip them. In particular, asymptotically AdS spacetimes become examples where the $C^{1,1}$ conformal extensions at infinity yield totally lightgeodesic boundaries, and all the results above apply. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:55:09 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Herrera",
"Jónatan",
""
],
[
"Sánchez",
"Miguel",
""
]
] |
2506.09033 | Haozhen Zhang | Haozhen Zhang, Tao Feng, Jiaxuan You | Router-R1: Teaching LLMs Multi-Round Routing and Aggregation via Reinforcement Learning | Code is available at https://github.com/ulab-uiuc/Router-R1 | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | The rapid emergence of diverse large language models (LLMs) has spurred the development of LLM routers that assign user queries to the most suitable model. However, existing LLM routers typically perform a single-round, one-to-one mapping (\textit{i.e.}, assigning each query to a single model in isolation), which limits their capability to tackle complex tasks that demand the complementary strengths of multiple LLMs. In this paper, we present \textbf{Router-R1}, a reinforcement learning (RL)-based framework that formulates multi-LLM routing and aggregation as a sequential decision process. Router-R1 instantiates the router itself as a capable LLM, leveraging its reasoning ability to interleave "think" actions (internal deliberation) with "route" actions (dynamic model invocation), and integrates each response into its evolving context. To guide learning, we employ a lightweight rule-based reward comprising format rewards, final outcome rewards, and a novel cost reward for performance and cost trade-off optimization, opening a pathway toward optimizing performance-cost tradeoffs via RL. Router-R1 also conditions only on simple model descriptors such as pricing, latency, and example performance, enabling strong generalization to unseen model selection. Experiments on seven general and multi-hop QA benchmarks show that Router-R1 outperforms over several strong baselines, achieving superior performance while maintaining robust generalization and cost management.Code is available at https://github.com/ulab-uiuc/Router-R1. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:56:45 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Zhang",
"Haozhen",
""
],
[
"Feng",
"Tao",
""
],
[
"You",
"Jiaxuan",
""
]
] |
2506.09034 | Sizhe Dang | Sizhe Dang, Yangyang Guo, Yanjun Zhao, Haishan Ye, Xiaodong Zheng, Guang Dai, Ivor Tsang | FZOO: Fast Zeroth-Order Optimizer for Fine-Tuning Large Language Models towards Adam-Scale Speed | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Fine-tuning large language models (LLMs) often faces GPU memory bottlenecks: the backward pass of first-order optimizers like Adam increases memory usage to more than 10 times the inference level (e.g., 633 GB for OPT-30B). Zeroth-order (ZO) optimizers avoid this cost by estimating gradients only from forward passes, yet existing methods like MeZO usually require many more steps to converge. Can this trade-off between speed and memory in ZO be fundamentally improved? Normalized-SGD demonstrates strong empirical performance with greater memory efficiency than Adam. In light of this, we introduce FZOO, a Fast Zeroth-Order Optimizer toward Adam-Scale Speed. FZOO reduces the total forward passes needed for convergence by employing batched one-sided estimates that adapt step sizes based on the standard deviation of batch losses. It also accelerates per-batch computation through the use of Rademacher random vector perturbations coupled with CUDA's parallel processing. Extensive experiments on diverse models, including RoBERTa-large, OPT (350M-66B), Phi-2, and Llama3, across 11 tasks validate FZOO's effectiveness. On average, FZOO outperforms MeZO by 3 percent in accuracy while requiring 3 times fewer forward passes. For RoBERTa-large, FZOO achieves average improvements of 5.6 percent in accuracy and an 18 times reduction in forward passes compared to MeZO, achieving convergence speeds comparable to Adam. We also provide theoretical analysis proving FZOO's formal equivalence to a normalized-SGD update rule and its convergence guarantees. FZOO integrates smoothly into PEFT techniques, enabling even larger memory savings. Overall, our results make single-GPU, high-speed, full-parameter fine-tuning practical and point toward future work on memory-efficient pre-training. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:56:53 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Dang",
"Sizhe",
""
],
[
"Guo",
"Yangyang",
""
],
[
"Zhao",
"Yanjun",
""
],
[
"Ye",
"Haishan",
""
],
[
"Zheng",
"Xiaodong",
""
],
[
"Dai",
"Guang",
""
],
[
"Tsang",
"Ivor",
""
]
] |
2506.09035 | Karhan Kayan | Karhan Kayan, Stamatis Alexandropoulos, Rishabh Jain, Yiming Zuo, Erich Liang, Jia Deng | Princeton365: A Diverse Dataset with Accurate Camera Pose | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce Princeton365, a large-scale diverse dataset of 365 videos with accurate camera pose. Our dataset bridges the gap between accuracy and data diversity in current SLAM benchmarks by introducing a novel ground truth collection framework that leverages calibration boards and a 360-camera. We collect indoor, outdoor, and object scanning videos with synchronized monocular and stereo RGB video outputs as well as IMU. We further propose a new scene scale-aware evaluation metric for SLAM based on the the optical flow induced by the camera pose estimation error. In contrast to the current metrics, our new metric allows for comparison between the performance of SLAM methods across scenes as opposed to existing metrics such as Average Trajectory Error (ATE), allowing researchers to analyze the failure modes of their methods. We also propose a challenging Novel View Synthesis benchmark that covers cases not covered by current NVS benchmarks, such as fully non-Lambertian scenes with 360-degree camera trajectories. Please visit https://princeton365.cs.princeton.edu for the dataset, code, videos, and submission. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:00 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Kayan",
"Karhan",
""
],
[
"Alexandropoulos",
"Stamatis",
""
],
[
"Jain",
"Rishabh",
""
],
[
"Zuo",
"Yiming",
""
],
[
"Liang",
"Erich",
""
],
[
"Deng",
"Jia",
""
]
] |
2506.09036 | Karthik Prabhu Palimar | Karthik Prabhu, Srinivasan Raghunathan, Ethan B. Anderes, Lloyd E. Knox | Learning Correlated Astrophysical Foregrounds with Denoising Diffusion Probabilistic Models | 21 pages, 11 figures, and 1 table; Our code and plotting scripts can be found on https://github.com/Karthikprabhu22/diffusion_model; to be submitted to JCAP; comments welcome | null | null | null | astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | Extragalactic foregrounds -- most notably the Cosmic Infrared Background (CIB) and the thermal Sunyaev-Zel'dovich (tSZ) effect -- exhibit complex, non-Gaussian structure and correlations that can bias analyses of small-scale cosmic microwave background (CMB) temperature anisotropies. These foregrounds can introduce mode coupling at small-scales (multipoles $\ell \geq 3000$) that mimic true lensing signals, thereby complicating analyses such as CMB lensing reconstruction. We present a novel approach to learn their full joint distribution using Denoising Diffusion Probabilistic Models (DDPMs) trained on paired CIB-tSZ patches at 150 GHz, from the Agora suite of extragalactic sky simulations. While simulations like Agora, which are based on N-body calculations, can take thousands of CPU hours, DDPM can synthesize realistic CIB-tSZ patches that faithfully reproduce both auto- and cross-spectral statistics of the 2-point, 3-point, and 4-point correlation functions, in a matter of seconds. We further demonstrate matching pixel-value histograms and Minkowski functionals, confirming that conventional non-Gaussian benchmarks are also satisfied. This framework provides a powerful generative tool for forward-modeling correlated extragalactic foregrounds in current and future CMB analyses. Although we mainly demonstrate the joint modeling of tSZ and CIB at a single frequency, we also include examples of its extension to multiple frequencies, showing that the framework can learn the spectral energy distributions (SEDs) across different bands. While establishing DDPMs as a promising tool for addressing foreground contamination in next-generation CMB surveys, we also outline remaining challenges to their practical deployment in analysis pipelines, such as scaling to larger sky areas and reliance on the underlying cosmological and astrophysical assumptions in the simulations used for training. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:00 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Prabhu",
"Karthik",
""
],
[
"Raghunathan",
"Srinivasan",
""
],
[
"Anderes",
"Ethan B.",
""
],
[
"Knox",
"Lloyd E.",
""
]
] |
2506.09037 | Matthew Ding | Matthew Ding and Robbie King and Bobak T. Kiani and Eric R. Anschuetz | Optimizing Sparse SYK | 27+7 pages, 1 figure. Abstract shortened to meet arxiv requirements | null | null | null | quant-ph cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Finding the ground state of strongly-interacting fermionic systems is often the prerequisite for fully understanding both quantum chemistry and condensed matter systems. The Sachdev--Ye--Kitaev (SYK) model is a representative example of such a system; it is particularly interesting not only due to the existence of efficient quantum algorithms preparing approximations to the ground state such as Hastings--O'Donnell (STOC 2022), but also known no-go results for many classical ansatzes in preparing low-energy states. However, this quantum-classical separation is known to \emph{not} persist when the SYK model is sufficiently sparsified, i.e., when terms in the model are discarded with probability $1-p$, where $p=\Theta(1/n^3)$ and $n$ is the system size. This raises the question of how robust the quantum and classical complexities of the SYK model are to sparsification.
In this work we initiate the study of the sparse SYK model where $p \in [\Theta(1/n^3),1]$. We show there indeed exists a certain robustness of sparsification. First, we prove that the quantum algorithm of Hastings--O'Donnell for $p=1$ still achieves a constant-factor approximation to the ground energy when $p\geq\Omega(\log n/n)$. Additionally, we prove that with high probability, Gaussian states cannot achieve better than a $O(\sqrt{\log n/pn})$-factor approximation to the true ground state energy of sparse SYK. This is done through a general classical circuit complexity lower-bound of $\Omega(pn^3)$ for any quantum state achieving a constant-factor approximation. Combined, these show a provable separation between classical algorithms outputting Gaussian states and efficient quantum algorithms for the goal of finding approximate sparse SYK ground states when $p \geq \Omega(\log n/n)$, extending the analogous $p=1$ result of Hastings--O'Donnell. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:08 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Ding",
"Matthew",
""
],
[
"King",
"Robbie",
""
],
[
"Kiani",
"Bobak T.",
""
],
[
"Anschuetz",
"Eric R.",
""
]
] |
2506.09038 | Mark Ibrahim | Polina Kirichenko, Mark Ibrahim, Kamalika Chaudhuri, Samuel J. Bell | AbstentionBench: Reasoning LLMs Fail on Unanswerable Questions | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | For Large Language Models (LLMs) to be reliably deployed in both everyday and high-stakes domains, knowing when not to answer is equally critical as answering correctly. Real-world user queries, which can be underspecified, ill-posed, or fundamentally unanswerable, require LLMs to reason about uncertainty and selectively abstain -- i.e., refuse to answer definitively. However, abstention remains understudied, without a systematic evaluation framework for modern LLMs. In this work, we introduce AbstentionBench, a large-scale benchmark for holistically evaluating abstention across 20 diverse datasets, including questions with unknown answers, underspecification, false premises, subjective interpretations, and outdated information. Evaluating 20 frontier LLMs reveals abstention is an unsolved problem, and one where scaling models is of little use. While recent reasoning LLMs have shown impressive results in complex problem solving, surprisingly, we find that reasoning fine-tuning degrades abstention (by $24\%$ on average), even for math and science domains on which reasoning models are explicitly trained. We find that while a carefully crafted system prompt can boost abstention in practice, it does not resolve models' fundamental inability to reason about uncertainty. We release AbstentionBench to foster research into advancing LLM reliability. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:30 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Kirichenko",
"Polina",
""
],
[
"Ibrahim",
"Mark",
""
],
[
"Chaudhuri",
"Kamalika",
""
],
[
"Bell",
"Samuel J.",
""
]
] |
2506.09039 | Abderrahime Filali | Abderrahime Filali, Diala Naboulsi, and Georges Kaddoum | Deep Reinforcement Learning-Based RAN Slicing with Efficient Inter-Slice Isolation in Tactical Wireless Networks | null | null | null | null | cs.NI | http://creativecommons.org/licenses/by/4.0/ | The next generation of tactical networks (TNs) is poised to further leverage the key enablers of 5G and beyond 5G (B5G) technology, such as radio access network (RAN) slicing and the open RAN (O-RAN) paradigm, to unlock multiple architectural options and opportunities for a wide range of innovative applications. RAN slicing and the O-RAN paradigm are considered game changers in TNs, where the former makes it possible to tailor user services to users requirements, and the latter brings openness and intelligence to the management of the RAN. In TNs, bandwidth scarcity requires a dynamic bandwidth slicing strategy. Although this type of strategy ensures efficient bandwidth utilization, it compromises RAN slicing isolation in terms of quality of service (QoS) performance. To deal with this challenge, we propose a deep reinforcement learning (DRL)-based RAN slicing mechanism that achieves a trade-off between efficient RAN bandwidth sharing and appropriate inter- and intra-slice isolation. The proposed mechanism performs bandwidth allocation in two stages. In the first stage, the bandwidth is allocated to the RAN slices. In the second stage, each slice partitions its bandwidth among its associated users. In both stages, the slicing operation is constrained by several considerations related to improving the QoS of slices and users that in turn foster inter- and intra-slice isolation. The proposed RAN slicing mechanism is based on DRL algorithms to perform the bandwidth sharing operation in each stage. We propose to deploy the mechanism in an O-RAN architecture and describe the O-RAN functional blocks and the main DRL model lifecycle management phases involved. We also develop three different implementations of the proposed mechanism, each based on a different DRL algorithm, and evaluate their performance against multiple baselines across various parameters. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:33 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Filali",
"Abderrahime",
""
],
[
"Naboulsi",
"Diala",
""
],
[
"Kaddoum",
"Georges",
""
]
] |
2506.09040 | Wei Song | Dianyi Wang, Wei Song, Yikun Wang, Siyuan Wang, Kaicheng Yu, Zhongyu Wei, Jiaqi Wang | Autoregressive Semantic Visual Reconstruction Helps VLMs Understand Better | null | null | null | null | cs.CV cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Typical large vision-language models (LVLMs) apply autoregressive supervision solely to textual sequences, without fully incorporating the visual modality into the learning process. This results in three key limitations: (1) an inability to utilize images without accompanying captions, (2) the risk that captions omit critical visual details, and (3) the challenge that certain vision-centric content cannot be adequately conveyed through text. As a result, current LVLMs often prioritize vision-to-language alignment while potentially overlooking fine-grained visual information. While some prior works have explored autoregressive image generation, effectively leveraging autoregressive visual supervision to enhance image understanding remains an open challenge. In this paper, we introduce Autoregressive Semantic Visual Reconstruction (ASVR), which enables joint learning of visual and textual modalities within a unified autoregressive framework. We show that autoregressively reconstructing the raw visual appearance of images does not enhance and may even impair multimodal understanding. In contrast, autoregressively reconstructing the semantic representation of images consistently improves comprehension. Notably, we find that even when models are given continuous image features as input, they can effectively reconstruct discrete semantic tokens, resulting in stable and consistent improvements across a wide range of multimodal understanding benchmarks. Our approach delivers significant performance gains across varying data scales (556k-2M) and types of LLM bacbones. Specifically, ASVR improves LLaVA-1.5 by 5% in average scores across 14 multimodal benchmarks. The code is available at https://github.com/AlenjandroWang/ASVR. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:57:50 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Wang",
"Dianyi",
""
],
[
"Song",
"Wei",
""
],
[
"Wang",
"Yikun",
""
],
[
"Wang",
"Siyuan",
""
],
[
"Yu",
"Kaicheng",
""
],
[
"Wei",
"Zhongyu",
""
],
[
"Wang",
"Jiaqi",
""
]
] |
2506.09041 | Avior Almoalem | Avior Almoalem, Rebecca Chan, Brinda Kuthanazhi, Juan Scmidt, Jose A. Moreno, Hermann Suderow, Paul Canfield, Taylor L. hughes and Vidya Madhavan | Discovery of a 1D edge mode in a Magnetic Topological semimetal | null | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | In rare-earth monopnictides like NdBi, the interplay between magnetism and topology results in an extremely unusual topological semimetal phase which simultaneously hosts Weyl points with Fermi arcs as well as massive and massless Dirac cones. A central question in this class of materials is whether ferromagnetic surfaces with gapped Dirac cones can also host robust well-defined chiral edge states. In this study, we use spin-polarized scanning tunneling microscopy (SP-STM) and spectroscopy to investigate the correlation between the magnetic and topological properties of NdBi. By combining SP-STM imaging with quasiparticle interference, we identify distinct signatures of both antiferromagnetic and ferromagnetic surface terminations and correlate them with their respective band structures. Crucially, we demonstrate that step edges on the ferromagnetic surface which serve as magnetic domain walls host well-defined one-dimensional (1D) edge modes that vanish above the N\'eel temperature. Our findings position NdBi as a promising platform for further explorations of 1D chiral edge modes and future realizations of Majorana states in proximitized rare-earth monopnictides. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:58:16 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Almoalem",
"Avior",
""
],
[
"Chan",
"Rebecca",
""
],
[
"Kuthanazhi",
"Brinda",
""
],
[
"Scmidt",
"Juan",
""
],
[
"Moreno",
"Jose A.",
""
],
[
"Suderow",
"Hermann",
""
],
[
"Canfield",
"Paul",
""
],
[
"hughes",
"Taylor L.",
""
],
[
"Madhavan",
"Vidya",
""
]
] |
2506.09043 | Tianyu Chen | Tianyu Chen, Darshal Shetty, Jeremy G. Siek, Chao-Hong Chen, Weixi Ma, Arnaud Venet, Rocky Liu | Gradual Metaprogramming | 13 pages, 10 figures | null | null | null | cs.PL | http://creativecommons.org/licenses/by/4.0/ | Data engineers increasingly use domain-specific languages (DSLs) to generate the code for data pipelines. Such DSLs are often embedded in Python. Unfortunately, there are challenges in debugging the generation of data pipelines: an error in a Python DSL script is often detected too late, after the execution of the script, and the source code location that triggers the error is hard to pinpoint.
In this paper, we focus on the F3 DSL of Meta (Facebook), which is a DSL embedded in Python (so it is dynamically-typed) to generate data pipeline description code that is statically-typed. We propose gradual metaprogramming to (1) provide a migration path toward statically typed DSLs, (2) immediately provide earlier detection of code generation type errors, and (3) report the source code location responsible for the type error. Gradual metaprogramming accomplishes this by type checking code fragments and incrementally performing runtime checks as they are spliced together. We define MetaGTLC, a metaprogramming calculus in which a gradually-typed metalanguage manipulates a statically-typed object language, and give semantics to it by translation to the cast calculus MetaCC. We prove that successful metaevaluation always generates a well-typed object program and mechanize the proof in Agda. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:58:32 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Chen",
"Tianyu",
""
],
[
"Shetty",
"Darshal",
""
],
[
"Siek",
"Jeremy G.",
""
],
[
"Chen",
"Chao-Hong",
""
],
[
"Ma",
"Weixi",
""
],
[
"Venet",
"Arnaud",
""
],
[
"Liu",
"Rocky",
""
]
] |
2506.09044 | Javier Sanguino | Javier Sanguino, Thomas Kehrenberg, Jose A. Lozano, Novi Quadrianto | The Decoupled Risk Landscape in Performative Prediction | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Performative Prediction addresses scenarios where deploying a model induces a distribution shift in the input data, such as individuals modifying their features and reapplying for a bank loan after rejection. Literature has had a theoretical perspective giving mathematical guarantees for convergence (either to the stable or optimal point). We believe that visualization of the loss landscape can complement this theoretical advances with practical insights. Therefore, (1) we introduce a simple decoupled risk visualization method inspired in the two-step process that performative prediction is. Our approach visualizes the risk landscape with respect to two parameter vectors: model parameters and data parameters. We use this method to propose new properties of the interest points, to examine how existing algorithms traverse the risk landscape and perform under more realistic conditions, including strategic classification with non-linear models. (2) Building on this decoupled risk visualization, we introduce a novel setting - extended Performative Prediction - which captures scenarios where the distribution reacts to a model different from the decision-making one, reflecting the reality that agents often lack full access to the deployed model. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 17:58:39 GMT"
}
] | 2025-06-11T00:00:00 | [
[
"Sanguino",
"Javier",
""
],
[
"Kehrenberg",
"Thomas",
""
],
[
"Lozano",
"Jose A.",
""
],
[
"Quadrianto",
"Novi",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.