Peer Revenue Play credits. Real papers.

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

All of arXiv

Thursday 8 October

Computer Vision

Towards Unified Evaluation of Prompt Enhancers for Video Generation

Yawen Shao, Yubo Zhu, Ziyun Dai et al. (+21 more)

Modern video generators can realize increasingly complex visual narratives, positioning the prompt enhancer (PE) as a critical bridge from concise user instructions and multimodal references to structured cinematic plans. However, existing PE evaluation relies on rendered videos, imposing substantial computational and human costs, slowing PE training and iteration, and conflating PE quality with downstream generator behavior. To address this gap, we introduce PEBench, the first unified benchmark for direct PE evaluation across text-to-video, image-to-video, and reference-to-video prompt enhancement. It comprises 1,100 expert-verified cases and 1,005 visual assets, spanning 35 fine-grained tasks with diverse temporal, cinematic, audiovisual, and multi-reference requirements. In addition, we develop PEBench evaluation, an evidence-grounded framework that combines modality-aware fact extraction with rubric-based assessment across 24 criteria. Our systematic evaluation of representative open- and closed-source PE methods reveals an emerging shift from fine-grained descriptive expansion toward intent-preserving cinematic planning, while the caption-reconstruction and forward-refinement methods show complementary strengths in cinematic coverage and semantic fidelity or internal coherence, respectively. Human validation shows that PEBench scores align closely with expert judgments of enhanced prompts and downstream videos from Wan3.0 and MiniMax-H3, indicating that prompt-level evaluation reliably reflects downstream utility.
Read the full abstract

All authors Yawen Shao, Yubo Zhu, Ziyun Dai, Zixun Fang, Kai Zhu, Zeyinzi Jiang, Yufeng Ai, Siyang Sun, Haolan Xue, Yu Shang, Yuxiang Bao, Zoubin Bi, Jingming Luo, Jie Xiao, Chaojie Mao, Zhehan Kan, Hongchen Luo, Yu Liu, Sheng Zhong, Wei Tong, Xueyang Fu, Yang Cao, Wei Zhai, Zheng-Jun Zha

Artificial Intelligence

Onboard Marine Anomaly Detection on $\Phi$sat-2: From Simulation-Based Development to In-Orbit Demonstration

Clotilde Szywala, Thomas Goudemant, Marjorie Bellizzi et al. (+2 more)

Onboard Artificial Intelligence can improve responsiveness and bandwidth efficiency of Earth Observation systems by processing data directly on the satellite. This paper presents the experience gained from the development, onboard integration, and post-launch adaptation of a lightweight marine anomaly detection pipeline deployed on the European Space Agency's $\Phi$sat-2 mission. The application combines sea segmentation, self-supervised feature encoding of marine regions, generic anomaly detection based on deviations from a normal sea state, and optional characterization of selected anomaly types. Before launch, the pipeline was trained and validated on simulated $\Phi$sat-2 imagery to assess algorithmic performance and compatibility with resource-constrained onboard hardware. After integration and functional validation in the mission environment, early experiments on real $\Phi$sat-2 acquisitions revealed a significant mismatch between simulated and in-orbit data. The pipeline was therefore retrained on real Level-1 imagery using an improved annotation strategy to better handle ambiguous marine regions, substantially enhancing performance. Beyond demonstrating the onboard feasibility of the application, the $\Phi$sat-2 experience highlights the importance of robust annotation strategies and sensor-aware design, and shows that simulation-based development is valuable for pre-flight risk reduction, while reliable scientific validation requires representative in-orbit data and should be clearly distinguished from functional validation.
Read the full abstract

All authors Clotilde Szywala, Thomas Goudemant, Marjorie Bellizzi, Benjamin Francesconi, Adrien Girard

Machine Learning

Timer-M1: A Multivariate Time Series Foundation Model via Learning Primitives

Haoran Zhang, Haixuan Liu, Xingjian Su et al. (+5 more)

We introduce Timer-M1, a pretrained multivariate time series foundation model that learns with primitives for zero-shot forecasting. Across domains, time series share elementary temporal and relational patterns, termed primitives, yet differ in how these primitives manifest and evolve across different contexts. Despite progress in zero-shot and task-general forecasting, existing foundation models may still struggle to generalize to complex real-world scenarios. To this end, we develop a primitive-based data synthesis and pretraining pipeline. The synthesis pipeline generates series with temporal primitives shared across domains and then assembles real and generated series into multivariate samples using relational primitives. Afterwards, samples are organized into episodes by assigning distinct channel roles as target variates, past-only covariates, and known-future covariates, ensuring that the model is optimized on predictable variates using available exogenous information. Technically, Timer-M1 further adapts gated two-dimensional Transformer blocks that dynamically allocate cross-variate attention across layers. Across three large-scale forecasting benchmarks, Timer-M1 ranks first on both FEV and TIME and second on GIFT-Eval among most recent time series foundation models. These results support effective primitive-based pretraining as a route to robust general forecasting technique across domains and task settings.
Read the full abstract

All authors Haoran Zhang, Haixuan Liu, Xingjian Su, Yong Liu, Zhi Chen, Yuxuan Wang, Jianmin Wang, Mingsheng Long

Materials Science

A General Algorithm for Minimal Uncoupled Stress–Strain Calculations of Elastic Constant

S. I. Simak and Utkarsh Singh

We have developed a general algorithm for constructing the minimum number of strain configurations required to extract second-order elastic constants from first-principles stress-strain calculations without algebraic coupling between the constants targeted for extraction. The method operates on the sparsity structure of the symmetry-reduced elastic stiffness matrix, represented by an auxiliary binary matrix, and generates admissible strain patterns algorithmically rather than prescribing separate strain sets for individual crystal symmetries. Each independent elastic constant is obtained from a single stress component, while additional stress relations generated by the same deformation provide internal consistency checks. Because the six-dimensional strain space contains only 2^6-1=63 non-empty strain patterns, the minimality of the resulting sets can be verified exactly for every elastic-symmetry type considered. The construction reproduces established high-efficiency strain sets in cases where they are minimal and yields improved sets for symmetry classes in which previously tabulated schemes require additional configurations. Because the formulation is based on stress-strain derivatives, it can also be readily extended to finite temperatures using consistently sampled ensemble-averaged thermodynamic stresses. Explicit strain sets are derived for all elastic-symmetry types spanning the 230 crystallographic space groups, and their accuracy and computational efficiency are benchmarked against established approaches.
Read the full abstract

Artificial Intelligence

MemTrial: Learning When to Trust Memory in LLM Portfolio Agents

Guanghao Wu, Zhuo Cai and Shoujin Wang

Large language model (LLM) agents for portfolio management learn from experience: they credit each experience in their memory with the outcome of the decisions that used it. In financial markets, however, this outcome mostly reflects the market move shared by all decisions on that date, so the credit tracks the market rather than the experience, and these agents often do worse than simply holding the equal-weight (1/$N$) portfolio. We ask how an agent can credit an experience with what it changes, and answer it by putting memory on trial: drafts of the same decision with and without an experience face the same market, so the outcome they share cancels in their difference. Our agent, MemTrial, drafts each decision with eight combinations of its retrieved experiences, chosen by a fractional factorial design, and credits each experience with its Banzhaf value, the average of these differences. As each date occurs once and each draft is a noisy LLM sample, these credits are noisy and may not hold on new dates. MemTrial therefore pools them across dates and similar experiences with a hierarchical Bayesian model, acts on them only after they have predicted unseen dates, and otherwise stays anchored at a conservative reference such as 1/$N$. On four benchmarks, MemTrial not only benefits from experiences that matter (the best of 15 methods on a semi-synthetic benchmark with known experience quality) but also limits its losses when its values do not hold (at most 2.2% below 1/$N$ on PortBench and InvestorBench, against 15–38% for the best experience-learning agent). Averaged over five settings, it improves the utility of the best experience-learning agent by 21.2%, and with eight LLMs it beats every LLM-based baseline on InvestorBench.
Read the full abstract

Materials Science

Interlayer ferromagnetic order revealed through pressure-enhanced interlayer exchange coupling in Fe$_3$GaTe$_2$

Riku Iimori and Takashi Kimura

Van der Waals ferromagnets possess finite interlayer magnetic coupling despite their layered crystal structures, and this coupling plays an important role in determining their magnetic properties. In this study, we performed magnetization measurements on high-quality Fe$_3$GaTe$_2$ single crystals while selectively tuning the interlayer coupling using hydrostatic pressure. We identified two characteristic temperatures, $T_{\rm inf1}$ and $T_{\rm inf2}$, below the Curie temperature with opposite pressure dependences: $T_{\rm inf1}$ is associated with thermally activated domain-wall depinning, whereas $T_{\rm inf2}$ reflects the formation of interlayer ferromagnetic order. Both temperatures exhibit pronounced pressure dependences, in sharp contrast to the nearly pressure-independent Curie temperature associated with intralayer ferromagnetic order. First-principles calculations revealed a substantial pressure-induced enhancement of the interlayer exchange interaction and reasonably reproduced $T_{\rm inf2}$ and its pressure dependence. These results demonstrate that the emergence of interlayer magnetic ordering can be predicted from the microscopic interlayer exchange coupling, establishing a predictive framework for understanding and designing interlayer coupling phenomena in layered magnetic materials.
Read the full abstract

Machine Learning

Spectral Weight Decay: Inducing Low-Rank Structure in Neural Network Weights

Dmitrii Andriianov, Andrey Veprikov and Aleksandr Beznosikov

Standard weight decay treats each weight matrix as a vector and ignores its spectral structure. We introduce spectral weight decay, a post-step decoupled nuclear-norm update that applies additive rather than multiplicative spectral shrinkage. We connect the update to approximate proximal descent and show that its sensitivity to update order can exceed that of conventional $\ell_2$ weight decay near rank deficiency. Across LLaMA models with $124$M to $500$M parameters, spectral weight decay lowers effective rank and improves SVD-LLM compression at matched validation loss. At $500$M and a $4\%$ distortion budget, it reaches $1.89\times$ compression and $1.18\times$ GPU inference speedup, compared with $1.14\times$ and $1.01\times$ after standard weight decay. Under fixed-horizon training with $60\%$ label noise, it also improves final mean clean-test accuracy over matched $\ell_2$ regularization by up to $17.8$ points on MNIST and $4.6$ points across four BERT-base tasks. Code is available at https://github.com/brain-lab-research/SpectralWD.
Read the full abstract

Biological Physics

Three-dimensional imaging of isolated membrane-protein complexes in vacuo with an X-ray laser

Kartik Ayyer, Tej Varma Yenupuri, Johan Bielecki et al. (+51 more)

The prospect of imaging single biomolecules, viruses and cells with intense, ultrashort X-ray pulses has driven the development of X-ray free-electron lasers (XFELs). However, the weak scattering from small particles is easily swamped by background from residual gas, which has so far limited applications to strongly scattering targets such as viruses, cell organelles and cells. Here we report a three-dimensional (3D) reconstruction of an isolated 1-MDa membrane-protein complex, photosystem I (PS I), from single-particle diffraction data. PS I trimers were aerosolised by charge-reduction electrospray ionisation and injected into the European XFEL beam, with partial helium gas exchange reducing background scattering by 80%. From 32 788 diffraction patterns of single trimers in random orientations, we reconstructed the 3D electron density to a resolution of 3.8 nm, limited by the detector geometry. The disc-shaped density, about 22 nm across and 10 nm thick, matches the size of a PS I trimer in a detergent micelle and is consistent with the compaction predicted by molecular dynamics simulations of the complex in vacuo and observed in native mass spectrometry. These results show that membrane-protein complexes can be imaged in vacuo with X-ray lasers, an important step towards ultrafast diffractive imaging of single macromolecules.
Read the full abstract

All authors Kartik Ayyer, Tej Varma Yenupuri, Johan Bielecki, Lutz A. Eichacker, J. Domingo Meza-Aguilar, Louise J. Persson, Lena Worbs, Patrick Adams, Roberto Alvarez, Adil Ansari, Alfredo Bellisario, Rebecca Boll, Carl Caleman, Sebastian Cardoch, Andrea Castoldi, Joe P. J. Chen, Cyril Danilevski, Merle Dechering, Simon Dold, Armando D. Estillore, Alberto De Fanis, Lukas V. Haas, Patrick Konold, David Lomitze, Abhishek Mall, Dimitra Manatou, Iñaki de Diego Martinez, Parichita Mazumder, Andrew J. Morgan, Anna Munke, Yevheniy Ovcharenko, Matteo Porro, Safi Rafie-Zinedine, Amit K. Samanta, Christina Schmidt, Arezu Sehati, Jonas A. Sellberg, Björn Senfftleben, Carolin Seuring, Zhou Shen, Nicusor Timneanu, Sergey Usenko, Daniel Westphal, August Wollter, Tamme Wollweber, Paul Lourdu Xavier, Henry N. Chapman, Janos Hajdu, Richard A. Kirian, Jochen Küpper, Erik G. Marklund, Petra Fromme, Michael Meyer, Filipe R. N. C. Maia

Combinatorics

Noncrossing combinatorics of quasi-Coxeter elements

Eugenio Cainelli, Giovanni Paolini and Riccardo Zaccaria

We study the combinatorics of the intervals below quasi-Coxeter elements in the absolute order of finite Coxeter groups. In particular, we prove that these intervals are EL-shellable, except possibly for quasi-Coxeter elements of types $E_7$ and $E_8$, and we compute their zeta polynomials. We also show that a total order of the reflections yields an EL-labeling if it does so on all rank-two subintervals, which proves a conjecture of Williams for Coxeter elements. In order to obtain these results, we prove them for the infinite families, and we treat the exceptional types by computer. In addition, we study the related interval monoids, and we find a counterexample to cancellativity for every proper quasi-Coxeter element of type $D_n$ ($n\ge5$), $E_6$, $E_7$, $E_8$; in particular, these monoids do not embed in the corresponding interval groups. This gives a negative answer to a question posed by Baumeister, Holt, Neaime, and Rees.
Read the full abstract

Software Engineering

MAP4CS: A Multi-dimensional Data Pruning Framework for Efficient Code Retriever Fine-tuning

Yuxuan Chen, Mingwei Liu, Guangsheng Ou et al. (+4 more)

Retrieval-Augmented Generation (RAG) has become a cornerstone in software engineering for enhancing Large Language Models (LLMs) with domain-specific knowledge. However, adapting retrievers to evolving code repositories remains challenging due to the noise and redundancy inherent in massive code corpora. Standard fine-tuning on the full corpus is computationally expensive and often leads to sub-optimal performance due to negative transfer from low-quality samples. Conversely, simple random sampling fails to guarantee data representativeness. To address these challenges, we propose MAP4CS (Multi-dimensional Awareness Pruning for Code Search), an adaptive data pruning framework. MAP4CS identifies a small, high-quality core subset by integrating syntactic structure, semantic diversity, and distributional representation, followed by a rigorous rule-based filtering pipeline. Extensive experiments on two large-scale datasets demonstrate that MAP4CS consistently outperforms random sampling baselines using only 5% of the training data. Remarkably, it achieves performance comparable to, or even superior to, fine-tuning on the full dataset, validating the ”less is more” hypothesis in data-centric AI. Furthermore, linguistic analysis reveals an adaptive optimization mechanism: MAP4CS automatically functions as a de-duplicator for redundant corpora and a denoiser for chaotic ones, constructing a training corpus that is both lexically diverse and information-dense.
Read the full abstract

All authors Yuxuan Chen, Mingwei Liu, Guangsheng Ou, Zekai Zhang, Zike Li, Yanlin Wang, Pelin Zheng

Differential Geometry

Rigidity of positively curved vacuum Static spaces

Gabjin Yun and Seungsu Hwang

In this paper, we study compact vacuum static spaces of dimension $\ge 4$ with nonnegative sectional curvature. First, we prove that an $n$-dimensional compact vacuum static space is isometric to a standard sphere provided it has positive sectional curvature and the complete divergence of its Weyl curvature tensor is nonnegative. Under the slightly weaker condition of nonnegative sectional curvature, we show that such a vacuum static space must have parallel Ricci curvature tensor. Second, we examine the critical point equation arising from the critical metrics of the total scalar curvature functional on the space of Riemannian metrics restricted to unit volume and constant scalar curvature. We demonstrate that a compact Riemannian manifold with nonnegative sectional curvature and nonnegative complete divergence of the Weyl curvature tensor, which admits a nontrivial solution to the critical point equation, must be Einstein and is isometric to a standard sphere. Our results can be viewed as extensions of the results in [4].
Read the full abstract

Software Engineering

Implementing the Spec Growth Engine: Preventing Spec-Code Divergence, and Growing the Spec with Agents

Hartwig Grabowski

The Spec Growth Engine anchors AI-assisted software development in a graph of specifications that the code is coupled to. This paper describes its implementation, which serves two tasks and keeps them apart as two layers. The first layer prevents spec-code divergence: a deterministic engine validates the spec graph, compares it with the code's import graph, earns a node's verified status from recorded test evidence, and classifies every change by what it can break – without calling a model. The second layer grows the spec with agents: an intent author, a planner and a coder, each played by its own model, extend the graph in rounds, and a deterministic rule decides after each round whether the run goes on. How much of the human's judgement is delegated is set by three independent switches – a draft gate, a delegation for breaking changes, and the run mode – which, with two ways of laying a project's floor, give eighteen ways to run a project. We describe each of them, the gates, requests and waivers through which agents and the human communicate, and the spectrum of operation from entirely manual work to an unsupervised run whose decisions the human reviews afterwards. Throughout, one claim holds the design together: an autonomous run is worth only as much as the deterministic instance that measures it.
Read the full abstract

Machine Learning

Addressing Overcommitment in the Reasoning of Gendered Economic Memes under Multimodal Ambiguity

Kushal Kanwar, Dushyant Singh Chauhan, Kapil Rana et al. (+2 more)

Multimodal meme understanding is increasingly used to analyze socially sensitive content, yet existing models often exhibit biased behavior when interpreting economic dependence and social roles under ambiguity. Many memes express economic relationships through sparse text or symbolic visual cues, providing insufficient evidence for gendered attribution. In such underspecified settings, models tend to rely on pretraining correlations, leading to hallucinated and stereotypical economic role assignments. In this work, we study gendered economic dependence in image-text memes through the lens of contextual sufficiency and identify epistemic overcommitment-inferring roles without adequate evidence-as a primary source of bias. We propose CGER-Net, a context-grounded multimodal framework that estimates whether the input provides sufficient evidence for gendered economic reasoning and applies evidence-gated inference to enable confident attribution when cues are explicit while favoring principled abstention otherwise. We evaluate CGER-Net on EconMeme-GE, a curated dataset of image-text memes annotated as Men, Women, Neutral, or Ambiguous. Across strong contemporary multimodal baselines, CGER-Net reduces Gender Overcommitment Rate by up to 44% on ambiguous instances while maintaining comparable accuracy on unambiguous cases. Human evaluation further shows that 79% of generated rationales are judged as epistemically aligned with the available evidence. These results highlight the importance of modeling when not to infer for reliable and responsible multimodal analysis.
Read the full abstract

All authors Kushal Kanwar, Dushyant Singh Chauhan, Kapil Rana, Gopendra Vikram Singh, Nils Lukas

Artificial Intelligence

MultiWorldBench: Do Independently Controlled Views Describe One Shared World?

Zhangbo Xu, Ruoxi Zhang, Rui Hu et al. (+3 more)

Multiplayer world models must ensure that independently controlled views remain consistent with a single shared and persistent world. We introduce MultiWorldBench, a diagnostic Minecraft benchmark comprising 495 case configurations across seven task suites and ten capabilities, including independent control, cross-view motion, shared-state synchronization, persistence, structural reasoning, concurrent interaction, and delayed revisit. Using Engine GT as a reference, we evaluate two multiplayer world models, Solaris (MP) and Gamma-World, together with a single-player generation variant, Solaris (SP). Among the generated systems, Gamma-World achieves the highest ten-capability average at 21.39, followed by Solaris (MP) at 20.88 and Solaris (SP) at 10.06, while Engine GT reaches 91.69. Gamma-World performs better on several control, shared-state, and revisit capabilities, whereas Solaris (MP) leads in cross-view motion and race-condition consistency. Nevertheless, all generated systems score at most 8.00 on state persistence and 1.33 on structural consistency, and none succeeds in spatial reasoning or building-identity preservation. Human pairwise evaluation yields the same overall system ranking and shows strong alignment with the automatic evaluation, with a mean dimension-level Spearman correlation of 0.95. These results show that current multiplayer world models still exhibit a substantial gap relative to Engine GT in maintaining a coherent shared world.
Read the full abstract

All authors Zhangbo Xu, Ruoxi Zhang, Rui Hu, Yisong Wang, Xitong Li, Jingdong Wang

Robotics

From Solo to Ensemble: A Hierarchical Framework for Composable Multi-Agent Human-Object Interaction

Zekai Deng, Kangyi Chen, Ye Shi et al. (+1 more)

Physics-based human-object interaction has achieved robust single-agent manipulation skills, yet extending them to multi-agent cooperative tasks remains challenging. Existing approaches typically adapt interaction policies through task-specific fine-tuning, which entangles low-level contact-rich execution with high-level coordination and limits reuse across object geometries, interaction types, and team sizes. We propose a hierarchical framework that converts a single-agent HOI policy into a reusable Object-oriented Motion Skill. Specifically, we reinterpret teacher rollouts as object-oriented action supervision by extracting short-horizon object-proxy motions from executed trajectories, and distill task-specific teachers into a low-level skill operating in an Object-oriented Action Space. For downstream tasks, the distilled skill is frozen as a reusable executor, while a high-level policy coordinates multiple agents by generating region-wise object-oriented actions conditioned on the shared object, task goal, agent states, and local manipulation regions. This formulation shifts multi-agent HOI learning from direct contact-rich full-body control to compact object-level proxy-motion coordination. Experiments on diverse HOI tasks show that the distilled Object-oriented Motion Skill supports robust proxy-motion execution and enables composable policy learning across different interaction types, object geometries, and team sizes.
Read the full abstract

All authors Zekai Deng, Kangyi Chen, Ye Shi, Jingya Wang

Graphics

SubDGuide: A Modeler-Inspired Agentic Workflow for Mesh-to-SubD Reconstruction

Mengnan Jiang, Christian Franke, Michele Franco Adesso et al. (+2 more)

Subdivision surfaces represent free-form geometry through a sparse control cage, but recovering that cage from a dense mesh is not merely a fitting problem: the system must infer where control is needed, which curves encode design features, and when an initial result should be revised. We present SubDGuide, a modeler-inspired agentic workflow for this task. Stage A reads aligned multiview geometric evidence and produces a compact plan for cage resolution, feature mapping, and broad-form fitting. Stage B inspects the resulting surface, requests targeted diagnostics, and chooses repair, rollback, or stopping actions. Geometry tools execute and verify every change; the planner never generates vertices or connectivity. On a fixed evaluation cohort, SubDGuide improves all five reported metrics over automatic remeshing: median Chamfer-L1 decreases from 0.641% to 0.443% of the target bounding-box diagonal, and F-score at a 1% tolerance increases from 82.00% to 92.78%. Stateful feedback improves four of five reported median metrics over one-shot planning, while verification retains the earlier checkpoint when a proposal is unhelpful. The same interface supports six multimodal planners, showing how semantic judgment can guide a practical, inspectable mesh-to-SubD workflow.
Read the full abstract

All authors Mengnan Jiang, Christian Franke, Michele Franco Adesso, Antonio Haas, Grace Li Zhang

Signal Processing

Outage-Aware Robust Sensing for UAV ISAC Systems

Qiming Li, Lei Zhang, Haoran Xu et al. (+1 more)

Integrated sensing and communications (ISAC) infrastructure can provide an external perception layer for aerial robots, but dynamic blockage makes the set of informative sensing links at the next update uncertain. A single predicted availability pattern can miss low-information outcomes, whereas protecting all patterns wastes power and may become infeasible when all links are blocked. We propose an outage-aware conformal robust power-allocation framework for multi-BS UAV tracking. A history-aware joint predictor maps radar and tracking histories to next-slot availability probabilities, and adaptive prediction sets retain plausible correlated patterns. Tracking constraints are enforced only for retained non-outage patterns, while the all-blocked pattern is treated separately as a physical sensing outage. With per-slot beam directions fixed, the resulting scalar-power allocation is a semidefinite program. A reliability bound relates serviceable-slot tracking failure to deployed-policy miscoverage and optimization infeasibility. Across paired closed-loop simulations over five random seeds, the method attains a 0.01% serviceable-slot constraint-violation rate and uses 47.0% less power than protection over all non-outage patterns.
Read the full abstract

All authors Qiming Li, Lei Zhang, Haoran Xu, Lina Mohjazi

Analysis of PDEs

Uniqueness for tv-regularized inverse problems with source in divergence form

L Baratchart, D P Hardin and C Villalobos-Guillén

Inverse source problems in divergence form consist in finding a vector field with prescribed support S, whose divergence is the Laplacian of some observed potential. In this paper, we assume the unknown vector field is a vector-valued measure, and we study the corresponding least square inversion problems, regularized by penalizing the total variation, without discretizing the criterion nor the unknown. We prove that this problem has a unique minimizer in the case where S is a slender set; i.e., it has zero Lebesgue measure and each connected component of its complement has infinite Lebesgue measure. The proof relies on a refinement of Smirnov's decomposition of divergence-free measures [62] which is of independent interest.
Read the full abstract

Optimization and Control

A Fast and High-Accuracy Finite-Time Sliding-Mode Algorithm for Computing Local Stackelberg Equilibria in Nonlinear Bilevel Games

Abdelhafid Zenati, Dave D Muir and Kamal Youcef-Toumi

Computing local Stackelberg equilibria in nonlinear sequential games is challenging because gradient-based best-response methods may diverge, cycle, or converge to stationary points that do not satisfy the required equilibrium conditions. These failures can result from curvature asymmetry, unstable response mappings, and lack of invariance with respect to the follower's best-response manifold. This paper introduces the Stackelberg Sliding-Mode Algorithm (SSMA), a control-theoretic framework that recasts Stackelberg equilibrium computation as a manifold-stabilization problem. The follower and leader first-order optimality conditions define Stackelberg sliding manifolds, and sliding-mode dynamics drive the iterates to these manifolds in finite time. The follower curvature and leader reduced-curvature conditions determine manifold attractivity, making valid local Stackelberg equilibria attractive while rejecting stationary points that fail the second-order equilibrium conditions. The resulting approach integrates finite-time manifold stabilization, Stackelberg-consistent dynamics, and curvature-based equilibrium selection within a unified framework. Across 500 nonconvex Monte Carlo trials, SSMA achieved 100% success, reached the prescribed tolerance in a median of 26 iterations, compared with 456 for the nearest fully successful baselines, and attained terminal residuals near \(10^{-16}\), while converging to a valid local Stackelberg equilibrium.
Read the full abstract

Formal Languages and Automata

Every Regular Language Has a Bipartite Cactus Automaton

Mario Grobler, Till Koziel and Sebastian Siebertz

Thompson's classical construction turns every regular expression into an equivalent epsilon-NFA whose transition graph is series-parallel. The construction uses epsilon-transitions, and eliminating them by the usual shortcut construction need not preserve planarity or the bound of two on the treewidth of the transition graph. Book and Chandra proved that every regular language nevertheless has an epsilon-free NFA with a planar transition graph. We show that the transition graph can in fact be chosen to be a bipartite cactus: every cycle is even, and any two cycles have at most one vertex in common. In particular, it is outerplanar and has treewidth at most two. This is a considerably more simple structure than the general series-parallel guarantee supplied by Thompson's construction. The proof is a direct induction on parity-typed regular expressions and uses only elementary closure properties of cactus graphs.
Read the full abstract

Machine Learning

4-Tensor Attention Model for Semantic Physical Reality

Jongwook Kim and Sangheon Yun

We describe a 4-tensor attention model that predicts the next semantic state of a scene, for video generation and robot planning. A window of states has positions (x, t) and two fibers, a semantic fiber and a temporal-context fiber, and one softmax normalizes attention jointly over the window. Frames and an agent's situation are written as those states; the encoder, the renderer, and the planner remain outside the update. To test the update on its own, we train on ROCStories, where each window poses the same next-sentence task at the semantic layer. On the validation split, with one seed per setting, the last-sentence cross-entropy on the three matched settings is lower for the 4-tensor model than for a free-running one-dimensional transformer by 5.3% at H=2, L=2, by 2.6% at H=4, L=2, and by 2.4% at H=4, L=3. At H=4, L=2 the parameter counts are nearly the same, 172.5M and 175.9M. On the same two GPUs that 4-tensor run finished in 2.4 hours and the baseline run in 45.2 hours; the baseline is trained by free-running decoding, one sequential forward pass per target token.
Read the full abstract

Artificial Intelligence

Internalizer: Portable Context-to-Parameter Mapping for Very Large Language Models

Peter Devine, Nick Ryan, Benjamin Sirb et al. (+1 more)

Hypernetworks that map a context directly to a LoRA adapter let a large language model carry that context in its weights, but prior work has demonstrated them only on base models of up to 14 billion parameters. We present the Internalizer, a state-of-the-art, portable Context-to-Parameter Mapping hypernetwork that generates document-specific LoRA adapters for the frozen 284B-parameter DeepSeek v4 Flash, a target two orders of magnitude larger than in any previous work. Most of its parameters live in a model-agnostic trunk with only thin entry and exit layers per base model, so it trains cheaply against small models before being ported to the large one. On unseen documents of up to 4096 tokens, the generated adapters reach 84.9% top-1 and 97.8% top-5 teacher-forced accuracy against 63.4% and 83.5% for the base model, with nothing in the context window but a three-word instruction. Once the hypernetwork is trained, a single forward pass turns any document into an adapter for such a model, which could be served alone for speed or alongside the document in the window to raise accuracy further.
Read the full abstract

All authors Peter Devine, Nick Ryan, Benjamin Sirb, Alex Chiocchi

Chemical Physics

Predictable Accuracy of Martini 3 Absolute Protein-Ligand Binding Free Energies across 15 Targets

Luis J. Walter and Tristan Bereau

Alchemical free-energy calculations predict protein-ligand binding affinities with close-to-experimental accuracy, but their cost restricts them to small sets of ligands. Coarse-grained (CG) models such as Martini 3 retain a rigorous statistical-mechanical treatment at a fraction of this cost, yet their accuracy for protein-ligand binding has not been assessed systematically. Here, we compute absolute binding free energies for 172 ligands across 15 protein targets with a Martini 3 double-decoupling protocol and compare them against atomistic FEP+ and OpenFE references. Martini underbinds for every target, by 2.7 to 8.7 kcal/mol. This offset is systematic: it comprises a uniform contribution that grows with the number of charged pocket residues. A linear model in this number, which requires only the protein structure, reduces the mean absolute error from 5.7 to 1.2 kcal/mol in leave-one-out cross-validation, close to the 1.0 kcal/mol obtained with exact per-target offsets. The ranking of ligands within a target varies more widely (median Spearman $\rho$ of 0.32, from $-$0.41 to 0.89) and is rationalized by the structure-activity relationship of each series: Martini ranks ligands well when their affinity is driven by hydrophobicity, $\pi$-stacking, or size, and poorly when it is driven by hydrogen bonding or charge. Both trends follow from Martini's calibration against bulk partitioning, which captures neither the dielectric environment of a binding pocket nor the directionality of hydrogen bonds. While MMGBSA retains the atomistic detail needed to rank congeneric ligands, Martini yields physically meaningful free energies, and its coarse-graining maps many atomistically distinct molecules onto one representation, making it a complementary tool for rapidly spanning chemical space.
Read the full abstract

Superconductivity

Electric and Heat Transport in One-Dimensional SNS Junctions with Energy-Dependent Transmission

Kai Tsutsumi and Shun Uchino

We investigate the phase-independent dc components of the electric and heat currents through a one-dimensional superconductor-normal-metal-superconductor (SNS) junction subject to chemical-potential and temperature differences between two superconducting reservoirs. Using a scattering approach based on the Bogoliubov-de Gennes equations, we incorporate multiple Andreev reflections while retaining the energy dependence of transmission through the normal region. We compare energy-independent, quantum-dot-type, and quantum-point-contact transmission profiles. At low bias, both currents exhibit nonlinear features associated with the superconducting gap and multiple Andreev reflections. Their responses to a temperature difference depend remarkably on the transmission profile. At a small finite bias, the electric current depends weakly on the temperature difference for the energy-independent and quantum-dot-type profiles, whereas the quantum-point-contact profile produces a pronounced nonlinear response. The heat current depends nonlinearly on the temperature difference for all three profiles over the parameter range investigated. These results highlight the role of the transmission profile in nonlinear electric and heat transport through SNS junctions and provide a microscopic reference for related transport studies in mesoscopic fermionic superfluid junctions.
Read the full abstract

General Relativity and Quantum Cosmology

Nonreciprocal black holes with primary hair in beyond Horndeski gravity: Exact solutions and stability

Sebastian Bahamonde, Theodoros Nakas and Simon Iteanu

We develop a novel systematic method for constructing exact, static, and asymptotically flat black-hole solutions in shift-symmetric and $Z_2$-symmetric beyond Horndeski theories. In contrast to previous studies, our approach does not rely on the restrictive reciprocal condition $g_{tt}g_{rr}=-1$, allowing for more general geometries characterized by two independent metric functions. The resulting solution space contains two qualitatively distinct types of hair originating from different branches of the theory. One branch is characterized by khronometric hair, while the other by genuine scalar hair. In both, the scalar profile defines a nontrivial preferred foliation, but the shift-symmetry Noether current and the associated propagating degree of freedom are trivial for the former and nonvanishing for the latter. Focusing mainly on the second branch, we derive several distinct families of exact black-hole solutions. In general, these solutions are singular and exhibit primary scalar hair associated with an independent scalar charge. Nevertheless, we demonstrate that they can be regularized to yield regular black-hole or solitonic configurations, for which the scalar hair becomes secondary. We further investigate the stability properties of these solutions under linear perturbations by analyzing the axial sector and the monopole $(\ell=0)$ mode of the even-parity sector. For representative solutions with scalar hair, the perturbative analysis provides evidence for linear stability under the modes considered.
Read the full abstract

Logic in Computer Science

Antichains for Concurrent Parameterized Games

Nathalie Bertrand, Patricia Bouyer and Gaëtan Staquet

Concurrent parameterized games involve a fixed yet arbitrary number of players. They are described by finite arenas in which the edges are labeled with languages that describe the possible move combinations leading from one vertex to another (n players yield a word of length n). Previous work showed that, when edge labels are regular languages, one can decide whether a distinguished player, called Eve, has a strategy to ensure a reachability objective, against any strategy profile of her arbitrarily many opponents. This decision problem is known to be PSPACE-complete. A basic ingredient in the PSPACE-membership proof is the reduction to the exponential-size knowledge game, a 2-player game that reflects the knowledge Eve has on the number of opponents. In this paper, we provide a symbolic approach, based on antichains, to compute Eve's winning region in the knowledge game. In words, it gives the minimal knowledge Eve needs at every vertex to win the concurrent parameterized reachability game. More precisely, we propose two fixed-point algorithms that compute, as an antichain, the maximal elements of the winning region for Eve in the knowledge game. We implemented these two algorithms in C++, as well as the one initially proposed, and report on their relative performances on various benchmarks.
Read the full abstract

Quantum Physics

Parameter-Dependent Noise Resilience in the Quantum Approximate Optimization Algorithm

Jorja Kirk and Ashley Montanaro

The Quantum Approximate Optimization Algorithm (QAOA) is a variational quantum algorithm used for solving combinatorial optimization problems, which is thought to be a promising application of near-term quantum hardware. However, high rates of noise in these devices can obscure meaningful results. Therefore, it is crucial to understand and suppress the effects of noise on the algorithm's output. Here, we find that the stability of the algorithm in the presence of noise is dependent on its variational parameters. Using a Pauli noise model, we derive an upper bound on the total variation distance between the noisy and noiseless circuit output as a function of these parameters. Results from noisy simulations support this bound and give insight into the best methods to use for parameter generation in the presence of noise. We also introduce new methods of optimization using error penalties and truncated search based on this bound to manufacture error resilience. We extend this analysis to the quantum adiabatic algorithm (QAA) by taking the continuous limit of the QAOA.
Read the full abstract

Functional Analysis

Uniform Approximation for Operator Learning on Compact Sets

Rui Liu and Jie Shen

We study uniform approximation for operator learning on compact subsets of Banach spaces. We bound the recovery dimensions and the number of network parameters in terms of compact-set complexity, the target gauge, and recovery stability. For bounded linear operators with a separable domain, we characterize gauge approximability by coordinate decompositions with a common vanishing gauge and arbitrarily small additive errors. Their partial sums give a fixed family of nonlinear encoder–decoder pairs. We prove that nonlinear recovery with a common vanishing gauge and a rank bound depending only on the model dimension is equivalent to the uniform approximation property (UAP). We also obtain linear recovery estimates with the same rank bound and an operator version for the linear factors of nonlinear targets. For Barron operators taking values in a Banach space $Y$ of type $p\in(1,2]$ and admitting a Bochner-integrable density, sampling gives the rate $N^{-(1-1/p)}$ for the expected $L^p(\lambda;Y)$ error, where $\lambda$ is a probability measure on the input compact set. For fixed finite-dimensional factors and a class of Hölder maps, we obtain matching upper and lower parameter bounds at fixed sufficiently large depth. Classical sampling and truncation examples provide explicit recovery maps, and a nonlinear integral operator illustrates the parameter bounds.
Read the full abstract

Commutative Algebra

Equivariant noncommutative crepant resolutions and $G$-maximal modifying modules

Yuki Hirano

Let $S$ be a Cohen-Macaulay normal domain containing a field $k$ with an action of a finite group $G$ with $|G|$ invertible in $k$. We prove that $G$-equivariant noncommutative crepant resolutions of $S$ induce noncommutative crepant resolutions of $S^G$ if the $G$-action is small. We also discuss similar results for equivariant maximal modification algebras. We introduce the notion of $G$-maximal modifying ($G$-MM) module, and we show that $G$-MM module over $S$ induces MM $S^G$-module in some special cases. Furthermore, we prove that Gorenstein MMAs of three-dimensional $\mathbb{Q}$-Gorenstein Cohen-Macaulay rings are derived equivalent. As an application of these results, we prove the existence of an MM module and derived equivalences of MMAs for three-dimensional terminal singularities of index two.
Read the full abstract

Artificial Intelligence

DeltaReplay: Task-Relative Memory Reuse for Mobile GUI Agents

Yudong Bai, Yihong Chen, Quanming Yao et al. (+1 more)

Memory-augmented mobile GUI agents store successful execution trajectories and reuse them in later tasks, but a stored trajectory rarely matches a new task exactly. The new task may use different parameters, share only some of its steps with a stored trajectory, or have no relevant record in memory. Forcing the agent to use irrelevant memory can mislead it, whereas discarding memory that may still be useful deprives it of guidance from past experience. To address this dilemma, we propose DeltaReplay, a step-level memory reuse framework that decides how to use existing memory without modifying it. We observe that the reusable part of a stored record is determined not by the record itself but by its relation to the new task, mainly through two factors: page-level consistency and action-level generality. We therefore store execution trajectories as paths in a transition graph, whose nodes (pages) and edges (actions between pages) capture these two factors. At reuse time, the action on each edge is split into a task-independent operation and task-specific parameters. DeltaReplay then compares each recorded step with the new task and the current screen, and decides whether to follow it, execute it after replacing its parameters, or leave it to the base agent. On AndroidWorld and SPA-Bench, DeltaReplay improves the task success rate over a base agent with the same backbone by up to 10.3 and 25.0 percentage points, respectively. These results indicate that deciding at each step how to use retrieved memory lets agents benefit even from partially matching trajectories.
Read the full abstract

All authors Yudong Bai, Yihong Chen, Quanming Yao, Yaqing Wang