Peer Revenue Publish or profit

Papers

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

All of arXiv

Sunday 4 October

Robotics

$R^2$-WAM: Repair-and-Reject Post-Training for World Action Models

Ruiyan Xu, Haisheng Su, Sixu Lin et al. (+4 more)

World Action Models (WAMs) emerge as a promising foundation for policy refinement by predicting the consequences of sampled actions. However, visually plausible predictions can mislead policy refinement if they fail to reflect the input actions. To address this mismatch, we introduce $R^2$-WAM, a two-stage repair-and-reject post-training framework that first improves the consistency of predicted futures with input actions, then uses these futures to select inferior action samples for negative fine-tuning. The repair stage grounds imagination in observed robot behavior through a kinematic alignment score that measures agreement between predicted and demonstrated motion, enabling the predicted video to faithfully reflect its input actions. Using the repaired video model, the rejection stage compares imagined outcomes of sampled and demonstrated actions, selectively applying negative fine-tuning to samples whose predicted task progress falls below the demonstrated reference by a prescribed margin. Together, the two stages extend video prediction from representation learning to consequence-based policy refinement without additional environment interaction or changes to the inference procedure. $R^2$-WAM achieves 93.8% average success on RoboTwin 2.0 across clean and randomized settings. On the long-horizon real-world Fold Shirt task, it achieves 87.5% average success, compared with 0% for Fast-WAM. Our project page is available at https://r2-wam.github.io/.
Read the full abstract

All authors Ruiyan Xu, Haisheng Su, Sixu Lin, Zhaokun Yue, Chengming Hu, Xin Jin, Guiliang Liu

Cryptography and Security

Static Bootstrap Placement for Encrypted Language Model Decoding

Halil Ibrahim Kanpak and Didem Unat

Language models increasingly serve prompts that carry private data, and secure inference under homomorphic encryption lets a client outsource the computation without revealing the prompt. Existing secure inference systems run a forward pass without consuming a token under encryption, and generating text with them requires a client round trip at every generated token. Keeping the loop on the server instead requires selecting and consuming a token under encryption, and placing bootstraps for a loop body that grows with the context. We build AR-HE, which runs the whole loop on the server, selects each token under encryption, retrieves its embedding, and writes it back into the encrypted state. The client sends one prompt and remains offline until the output. One rule places every bootstrap in the run, without search, so the bootstrap cost of a token is a formula in the context length that is known before the run starts. The schedule skips work whose result cannot reach the output, packs bootstraps that share an operand, and keeps the keys and values of past positions in an encrypted cache. With every optimization applied, generating a GPT-2 small token costs 544 seconds on one NVIDIA H100, down from 4715 seconds without optimization. The prompt step before it costs 4630 seconds. The cache alone takes a generated step from 11751 bootstraps to 1072. The formula predicts every step we measured, including steps of a model it was not derived from.
Read the full abstract

Artificial Intelligence

VideoResearchAgent: Grounded Task Synthesis and Sim-to-Real RL for Open-Web Video Research

Yuhang Zhou, Fei Li, Yuxi Wu et al. (+2 more)

Existing deep research agents are designed primarily for text- and image-based web sources, while video reasoning systems typically assume that relevant videos are provided in advance. We study open-web video research, where an agent must autonomously discover relevant videos, navigate their temporal content, and ground answers in visual evidence. Training such agents at scale is challenging as live video interaction is slow and unreliable, whereas fixed local simulation can induce retrieval-specific shortcuts that fail to transfer to the open web. We introduce VideoResearchAgent, a scalable training framework to address these challenges. First, we introduce controllable task synthesis pipeline to synthesize multi-hop research tasks from timestamped visual evidence while filtering text-only shortcuts. Second, we build a field-aligned local video simulator that preserves deployment-facing search and watch interactions while accelerating video search by a factor of 34.5-64.6. Third, we introduce Retrieval-Domain-Randomized GRPO (RDR-GRPO), which diversifies candidate rankings, distractors, metadata, and result structure during training to reduce overfitting to simulated retrieval. On Video-BrowseComp, the VideoResearchAgent trained using Qwen3.5-4B achieves 40.48% accuracy, comparable to Gemini-3-Flash-Preview, while reducing cumulative API-token consumption by 74.9% relative to the untrained model. Together, these results establish an accurate and efficient training recipe for open-web video research.
Read the full abstract

All authors Yuhang Zhou, Fei Li, Yuxi Wu, Bin Zhu, Jingjing Chen

Robotics

GJK-CBF: Control Barrier Functions for Convex Rigid Body Collision Avoidance on SE(3)

Yi-Hsuan Chen, Shuo Liu, Wei Xiao et al. (+2 more)

Collision avoidance among convex bodies is a fundamental problem in robotics. Control Barrier Functions (CBFs) provide a practical framework for real-time safety filtering due to their computational efficiency. For general convex bodies, exact separation measures, such as distance or scaling factor, are typically computed through optimization. Existing CBF formulations often obtain the required gradient via differentiable optimization (diffOpt), adding computational overhead. In contrast, we leverage the Gilbert-Johnson-Keerthi (GJK) algorithm to obtain the current witness pair—the pair of points realizing the minimum distance or penetration depth—and formulate a CBF, termed GJK-CBF, whose gradient is constructed directly from the relative rigid-body motion of the witness pair, without resorting to diffOpt. This formulation applies to both 2D and 3D environments across a broad class of convex body pairs, provided that at least one in each one-to-one interaction is strictly convex. The proposed GJK-CBF is validated in various scenarios, including multi-robot position swapping in both 2D and 3D, navigation through a vertical slit in 3D, and its applicability to manipulators. The results demonstrate collision-free motion across all scenarios while reducing the conservativeness introduced by geometric approximations, particularly in narrow environments.
Read the full abstract

All authors Yi-Hsuan Chen, Shuo Liu, Wei Xiao, Michael Otte, Calin Belta

High Energy Physics (Theory)

The Phenomenology of the Dark Dimension and Non-Supersymmetric Strings

Cumrun Vafa

Swampland principles, combined with the smallness of dark energy $\Lambda$, lead to the dark dimension scenario, with one extra dimension of micron size. We review this scenario and its predictions: KK graviton dark matter (or the KK tower of the quintessence field into which the gravitons convert), neutrino masses tied to $\Lambda^{1/4}$ via a bulk $B-L$ gauge symmetry, QCD axion decay constant $f_a\sim \Lambda^{1/12}\sim 10^9-10^{10}$ GeV at the 5d Planck scale $M_5$, connecting electroweak hierarchy to $\Lambda^{1/6}$, and a relation between the GUT and weak scales with $M_5$ as the geometric mean. The dark dimension also provides a concrete realization of an evolving dark sector: the evolution of its diameter correlates the variation of dark energy with that of the dark matter mass, fitting the recent DESI DR2 (+CMB+supernovae) data while avoiding the phantom crossing that the $w_0w_a$CDM parametrization would suggest—the apparent phantom behavior instead arises from the evolving dark matter mass. The fits suggest that the dark energy field sits in the interior of moduli space, near an unstable critical point. Motivated by this, we review recent progress on non-supersymmetric string vacua: isolated critical points protected by duality symmetries, in non-geometric flux compactifications and in Scherk–Schwarz compactifications of M-theory, and an emerging duality web for non-supersymmetric strings. This note is based on two talks given by the author at String Phenomenology 2026 (Paris) and Strings 2026 (Shanghai).
Read the full abstract

Quantum Physics

Custom Penalty QAOA via Walsh-Hadamard Transform

Xin Wei LEE, Siong Thye GOH and Hoong Chuin LAU

Inequality constraints in combinatorial optimization are conventionally handled by introducing slack variables and quadratic penalties, which increase the qubit count and distort the optimization landscape. Slack-free formulations instead apply a nonlinear custom penalty function directly to the constraint, which are difficult to be implemented in quantum approximate optimization algorithm (QAOA). We show that the Walsh–Hadamard transform (WHT) removes this obstruction for any penalty function: a penalty applied to a linear constraint is exactly a weighted sum of Pauli-$Z$ strings. We propose this framework as the Walsh custom penalty QAOA (WCP-QAOA). Truncating the expansion at Walsh degree $k$ costs $\mathcal{O}(n^k)$ Pauli terms and leaves the qubit count untouched. The truncation is the optimal degree-$k$ approximation of the penalty in $L^2$, and we give a condition on the discarded Walsh coefficients under which the truncated Hamiltonian still ranks every feasible solution below every infeasible one. Since evaluating a Walsh coefficient from its definition requires all $2^n$ penalty values, we give an algorithm that returns any coefficient in sub-exponential time, so that the degree-$k$ Hamiltonian is obtained in time polynomial in $n$ for fixed $k$. We show that for the exponential penalty, there exists a closed product form of WHT, and hence a closed form expression to calculate the Walsh coefficients. We identify structural shortcuts that shorten the exact spectrum. We also tested WCP-QAOA in solving a 6-qubit knapsack problem.
Read the full abstract

Machine Learning

Why Subliminal Learning Needs So Much Data: A Noisy Inverse View through Steering Vector Recovery

Luoyu Chen, Xiaoyu Ding, Weiqi Wang et al. (+4 more)

Subliminal learning lets a student inherit a teacher's behavioral trait from semantically unrelated data, yet published demonstrations typically require tens of thousands of carrier examples. We ask where this data requirement comes from. Our testbed is subliminal steering: the teacher trait is a known residual-stream vector $\Delta_T$, so transfer can be measured directly as parameter recovery. On identical carrier prefixes, we compare token-level (hard) NLL supervision with full-distribution (soft) KL supervision. At initialization the two objectives give nearly collinear gradients, and both align poorly with $\Delta_T$. Under iterative optimization, however, they diverge: soft supervision recovers $\Delta_T$ almost exactly from a few hundred carriers, while hard supervision stays well below it even with tens of thousands. We explain this gap by casting steering-vector distillation as a noisy linear inverse problem. Locally, the carrier task maps the trait through its Fisher matrix $F$, so gradients point toward $F\Delta_T$ rather than $\Delta_T$. Gradient descent then acts as a progressively less-damped inverse of $F$. With soft targets, this inverse restores low-curvature directions. With hard labels, it also amplifies the sampling noise in those same directions. The result is an optimal inversion depth that grows with the number of independent carriers. Experiments on Qwen2.5-7B and Gemma-2-9B confirm four predictions: the Fisher distortion of the initial gradient, recovery ordered from steep to flat directions, an optimal depth that shifts with data scale, and the finding that resampling completions from a fixed prompt pool works as well as adding new prompts. In this setting, large carrier datasets are needed less to reveal the trait than to suppress label noise amplified by Fisher inversion. Code is available at \url{https://github.com/luoyuchenmlcv/subliminal-data}.
Read the full abstract

All authors Luoyu Chen, Xiaoyu Ding, Weiqi Wang, Chenhan Zhang, Zhiyi Tian, Jianhuan Huang, Shui Yu

Computation and Language

Scaling Verifiable Environments for Long-horizon Work Agents

Jiazheng Zhang, Long Ma, Yunxian Yang et al. (+23 more)

Work agents operate over digital artifacts to execute professional knowledge-intensive work, requiring training environments that support long-horizon interaction and trustworthy verification. However, hand-crafted environments incur prohibitive engineering overhead that prevents environment scaling, whereas synthesis methods sacrifice workspace complexity, realism, or grounded verifiability. To bridge this gap, we introduce WorkForge, a scalable synthesis framework for constructing verifiable work-agent environments from real-world resources. Starting from expert workflows, WorkForge first identifies the resources, decisions, and deliverables required by each workflow. It then retrieves relevant real-world files and organizes them into a workspace. WorkForge inspects the workspace to extract concrete, checkable facts about its content. These factual anchors fix which task types the workspace can support and how their outcomes can be verified. Therefore, WorkForge derives each task's instructions, solution plan, and complementary programmatic and semantic verifiers directly from these factual anchors, keeping verification traceable to observable workspace evidence. Furthermore, we construct 16.7K verifiable environments across 40 professional domains, with workspaces collectively covering 60 file types. Post-training Qwen3.5-35B-A3B-Base improves GDPVal from 45.5 to 73.6 and APEX Score from 5.0 to 21.3, while enabling Qwen3.5-27B to achieve highly competitive performance and outperform strong competitors. Our analyses confirm the efficacy of the proposed method and reveal consistent scaling behaviors across both data volume and interaction horizons.
Read the full abstract

All authors Jiazheng Zhang, Long Ma, Yunxian Yang, Zhiheng Xi, Zhikai Lei, Yajie Yang, Chenyang Liao, Enyu Zhou, Yang Nan, Yuchen Tian, Senjie Jin, Yibo Wang, Wei He, Boyang Liu, Jixuan Huang, Xin Guo, Shutong Jiang, Zhezheng Hao, Xinbing Liang, Zhihao Zhang, Changzhi Zhou, Wiggin Zhou, Tao Gui, Qi Zhang, Xuanjing Huang, Clarenceai

Machine Learning

ScopeSAE: Model-Scope Feature Discovery with Interpretable Layer Selection

Qingwen Zeng, Zehao Fu, Shuyu Meng et al. (+5 more)

Sparse autoencoders (SAEs) are a central tool in mechanistic interpretability. However, existing SAEs are primarily trained per layer. The modeling subspace is therefore fixed by layer identity, independent of which token-layer states actually drive each prediction. We argue that this constraint contributes to several limitations observed in layer-wise SAEs, including low feature utilization, high dictionary redundancy, and features that lack direct behavioral grounding. In this paper, we propose ScopeSAE, which selects the modeling subspace per token by attributing each prediction to its most influential token-layer state via normalized gradient-based attribution, and learns features over the resulting prediction-relevant subspace. Empirically, ScopeSAE yields an effect we term reconstruction-better-than-original. Written-back reconstructions of the SAE produce lower next-token cross-entropy than the original activations, an outcome that, to our knowledge, has not previously been reported for SAEs. Through interventional analyses and a KL fine-tuning counter-experiment, we show that this effect is attributable to ScopeSAE's prediction-relevant subspace itself rather than to architectural changes. ScopeSAE further improves effective feature count, interpretability, utilization, and dictionary redundancy over existing layer-based baselines, suggesting that choosing the SAE modeling subspace by predictive relevance leads to more useful and behaviorally meaningful features.
Read the full abstract

All authors Qingwen Zeng, Zehao Fu, Shuyu Meng, Linghan Huang, Jiayi Zhang, Chenglin Wu, Ling Chen, Huaming Chen

Information Retrieval

ModelLakeFishing: Efficient Retrieval over Million-Scale Model Lakes

Xiaoyang Liu, Zhengyuan Dong and Renée J. Miller

Open model lakes may contain millions of reusable models, making it costly to identify suitable models for a new dataset. We present ModelLakeFishing, a model-retrieval framework for queries specifying a target dataset, prediction task, and evaluation metric. It consolidates metadata and historical evaluations into a model-dataset-task evidence graph, learns model and query embeddings with a relation-aware graph encoder, and indexes model embeddings using Hierarchical Navigable Small World (HNSW) search. At query time, HNSW retrieves 1,000 candidates without scoring every model, after which a training-side task prior reranks candidates for the requested metric and returns the top 10. We evaluate on a lake of 3,016,439 models and 247,803 observed model-dataset performance pairs using three root-aware splits that hold test performance edges out of representation learning and retrieval. ModelLakeFishing achieves a mean eligible-query gold@10 of 0.2968, recovering the observed-best model in the top 10 for 29.68% of eligible queries and retaining 93.47% of the gold@10 of an exhaustive baseline using the same scoring and reranking procedure. Given precomputed query embeddings, retrieval and reranking take 0.747 ms median and 1.102 ms at the 95th percentile. These results demonstrate efficient retrieval over million-model lakes from sparse relational evidence.
Read the full abstract

Materials Science

Anisotropic mesoscale magnetic scattering across the spin-reorientation transition in TbMn6Sn6

Zhuang Xu, Fan Chen, Hui Cheng et al. (+10 more)

TbMn6Sn6 is a centrosymmetric ferrimagnetic kagome metal that combines topological electronic states with strong magnetic anisotropy and a near-room-temperature spin-reorientation transition (SRT). Real-space imaging studies of thin specimens have revealed helical stripes, biskyrmion-like textures, and skyrmion-bubble-like structures in the spin-reorientation regime. Whether such local textures develop into a bulk long-range-ordered skyrmion lattice remains unresolved. Here we present small-angle neutron scattering (SANS) results probing the mesoscale magnetism of a bulk TbMn6Sn6 single crystal. We observe no sixfold SANS Bragg pattern indicative of a long-range-ordered skyrmion lattice under the explored conditions. Instead, the dominant magnetic response is an anisotropic twofold low-Q scattering component whose intensity distribution and field dependence evolve markedly across the spin-reorientation regime. In zero field, a broad and horizontally elongated low-Q signal develops above TSR and weakens as the system approaches the paramagnetic state. At 345 K with H || b, the signal is diffuse and is suppressed rapidly below about 700 Oe. By contrast, at 310 K with H || c, close to TSR, the response is more concentrated and line-shaped and persists to higher fields. The contrasts between these scans likely reflect the combined effects of rapidly evolving magnetic anisotropy near TSR and the field geometry. Our results therefore establish a bulk anisotropic mesoscale magnetic response across the spin-reorientation regime, consistent with stripe- or domain-related textures reported by real-space imaging, while not uniquely identifying their real-space morphology.
Read the full abstract

All authors Zhuang Xu, Fan Chen, Hui Cheng, Xinmin Wang, Taisen Zuo, He Cheng, Haitao Hu, Cuiping Zhang, Qian Zhao, Michel Kenzelmann, Jonathan S. White, Junying Shen, Xin Tong

Artificial Intelligence

Self-Evaluating Recursive Agents

TianYi Lyu, Xiaozhe Li, Yang Li et al. (+7 more)

Recursive language-model agents decompose tasks and delegate subtasks to child instances of the same policy, forming a tree of work. Training them, however, is hard: the final outcome is verifiable, but the self-invented intermediate subtasks are numerous and carry no ground truth. Existing methods score each node with a verifier or judge, which is costly at scale and blind to decomposition quality. We argue that a recursive agent must learn three coupled capabilities within one set of weights: decomposing problems into subtasks, solving them, and evaluating the outcomes, each requiring its own training signal. SERA (Self-Evaluating Recursive Agents) turns evaluation into a learned capability of the policy itself. Before delegating, the parent writes a rubric of weighted success criteria for each child subtask; a ranking objective against verified outcomes then trains rubric generation so that the criteria track genuine subtask success. In addition, a complementary leaf-coverage signal provides direct credit for task decomposition. Our central finding is that training the policy to generate aligned rubrics is what drives the gains: because the same weights both evaluate and execute, learning to judge subtasks sharpens the agent's ability to solve them. Notably, external supervision is also reduced: the judge is consulted only to train the rubric generator, while solving is trained against the agent's own rubric scores, which outperform direct use of the judge. Beyond training, the learned rubric doubles as an inference-time selector for tree search. On TextCraft-Synth and TextWorld-Sync, SERA improves over strong recursive-agent baselines by 5.38 and 13.14 points on average, and rubric-guided tree search at inference adds a further 2.43 points on TextWorld-Sync.
Read the full abstract

All authors TianYi Lyu, Xiaozhe Li, Yang Li, Yongkang Chen, Kefei Tian, Junbo Niu, Zican Hu, Hongbo Liu, Mingliang Xiong, Qingwen Liu

Analysis of PDEs

Long-Time Nonlinear Stability of Large-Amplitude Radial Rarefactive Flows for the Multidimensional Compressible Euler Equations

Faris El-Katri

In this paper, we study multidimensional perturbations of radially symmetric rarefactive flows for the compressible Euler equations. No assumptions on the smallness of the background are imposed. We derive lower bounds on both the classical lifespan and stability timespan of the perturbed solution. These lower bounds are demonstrated to tend to infinity as the size of the perturbation approaches zero. Furthermore, owing to the rarefactive nature of the background, the resulting estimate is shown to be substantially longer than the corresponding estimate for a general classical background with bounded first derivatives.
Read the full abstract

Systems and Control

A Unified Framework for Robust Data-Driven Model Predictive Control with Terminal Ingredients

Zhaohua Yang, Nan Bai, Pengyu Wang et al. (+1 more)

This paper proposes a unified framework for robust data-driven model predictive control (MPC) with terminal ingredients for unknown linear time-invariant systems. Based on Willems' fundamental lemma, a single offline input-output trajectory is used as an implicit model, while online implementation only requires past input-output measurements. Bounded measurement noise is considered in both offline and online output data. The framework consists of an offline stage, where robust terminal ingredients are designed directly from noisy data, and an online stage, where these ingredients are incorporated into a one-step MPC scheme. Recursive feasibility of the proposed framework and practical exponential stability of the closed-loop system are established. Numerical results illustrate the performance tuning capability of the terminal set, with particularly pronounced benefits for short prediction horizons, and show that the proposed framework achieves performance comparable to data-driven MPC without terminal ingredients while requiring a reduced prediction horizon and less online computation.
Read the full abstract

All authors Zhaohua Yang, Nan Bai, Pengyu Wang, Ling Shi

Computation and Language

Rewrite What Matters: Adaptive Multilingual Query Rewriting for Reasoning via Agentic Reinforcement Learning

Rui Qi, Yufeng Chen, Yunlong Liang et al. (+6 more)

In multilingual scenarios, queries with equivalent semantics but in different languages could guide the model into different reasoning trajectories, leading to performance disparities. To mitigate this gap, previous studies typically apply a one-size-fits-all query rewriting strategy, such as translation, which overlooks the fact that different scenarios require diverse types of semantic transformations. In this paper, we propose mRewriter-R1, an agentic multilingual query rewriting framework with reinforcement learning. Unlike single-turn rewriting, mRewriter-R1 formulates multilingual query rewriting as a multi-turn sequential decision-making process, where the model dynamically performs multi-aspect optimization through adaptive operator selection. Experimental results demonstrate that mRewriter-R1 outperforms all strong multilingual rewriting baselines on different large reasoning backbones. Further analyses show that the learned policy can adaptively decide on rewriting operators according to query characteristics, exhibiting strong generalization ability across diverse reasoning tasks, and plug-and-play compatibility with heterogeneous reasoning language models.
Read the full abstract

All authors Rui Qi, Yufeng Chen, Yunlong Liang, Chuan Meng, Sijin Lu, Ge Shi, Jinan Xu, Fandong Meng, Kaiyu Huang

Computation and Language

A Systematic Analysis of the Predictive Power of LM Surprisal in Reading Chinese

Hongao Zhu, Muxiaoqiao Xu, Yikang Liu et al. (+4 more)

This study analyzes the predictive power of LM-derived, token-level surprisal on Mandarin Chinese reading times. Using a suite of Chinese-Pythia models (14M–1.4B) trained from scratch with 30B tokens, we examine how well surprisal predicts first fixation duration, gaze duration, and total reading time in three paragraph-level eye-tracking corpora of Mandarin Chinese (GECO-CN, HKP, and MECO). Contrary to previous null findings, our results show that surprisal is predictive of Chinese reading times. However, whether predictive power scales with model size and the amount of training is corpus-specific: bigger models predict reading times better in GECO-CN, whereas inverse scaling emerges in HKP and, at the largest sizes, in MECO. Subsequently, we find that checkpoints whose surprisal remains closer to $n$-gram statistics are better predictors of reading in HKP and MECO, providing one possible explanation for the inverse scaling behavior. All in all, the predictive power of surprisal on Chinese reading time measurements is corpus-specific, which cautions against drawing conclusions about language model scale from a single corpus.
Read the full abstract

All authors Hongao Zhu, Muxiaoqiao Xu, Yikang Liu, Siyuan Song, Yuxia Wang, Byung-Doh Oh, Hai Hu

Cryptography and Security

WLPA: A Network Management Framework for Allocating Scarce Quantum-Safe Link Postures under Weakest-Link Exposure

Bhanwar Gupta and Sanjeev Rana

Network operators migrating distributed systems to post-quantum and quantum-safe cryptography face a resource-management problem: quantum key distribution (QKD) gives the strongest guarantee available but is hardware-scarce and costly, so only a fraction of a deployment's links can receive it. Default practice – ranking links by traffic or centrality and upgrading the top scorers – optimizes each link in isolation, while a distributed job's real security is set by whichever active link ends up weakest. We present WLPA (Weakest-Link Posture Assignment), an implementation-ready decision framework and algorithm for allocating scarce quantum-safe postures under hardware, latency, and compliance constraints. WLPA computes the optimal weakest-link posture in closed form and returns an assignment in O(|E| log |E|) time – 125 ms for a 20,000-link network on a laptop – with a breach-aware refinement matching an exact integer program under common conditions. Our central contribution is a dichotomy telling operators exactly when existing per-link heuristics already match this optimum and when they systematically fail: under constant upgrade cost and uniform traffic-driven risk the two coincide exactly; once costs vary or an adversary concentrates on weak targets, heuristics leave the weakest link unprotected, unboundedly. We validate WLPA across 19 experiments: four microservice topologies, a real enterprise network (SNAP email-Eu-core), a post-quantum latency benchmark, a Qiskit BB84 simulation and a real-hardware BB84 run on IBM's ibm_marrakesh processor, checks against real production traffic traces, a CVE-calibrated correlated-compromise study, Bayesian uncertainty quantification, and seven baseline methods. A concrete-security reduction grounds the framework as theoretical foundation, not primary contribution.
Read the full abstract

Robotics

Tackling Sim-to-Real Mismatch Through Sampling-Based Disturbance Observers: From Analytical Models to Learned World Models

Tianqi Zhu, Jun Yang, Jianliang Mao et al. (+2 more)

Robotic controllers increasingly rely on analytical models, simulators, cost-query interfaces, and learned world models. However, physical deployment can deviate from nominal assumptions, and additional disturbances may arise even when the model itself is accurate. In control systems, disturbance observers (DOB) are widely used to estimate such unmeasured effects from nominal models and measured feedback. Classical DOB formulations are generally built around explicit plant models. This paper develops the sampling-based disturbance observer (SDOB), extending the DOB principle to a broader range of models, including simulators and learned world models, through state-rollout or cost-query interfaces. SDOB separates two observable channels: state-effect disturbances, for which the feedback state differs from its prediction, and cost disturbances, for which the same query state receives different costs as the perceived environment changes. Diverse simulation and real-robot experiments across traditional and learned models demonstrate the effectiveness of SDOB in compensating for sim-to-real mismatch and improving control performance.
Read the full abstract

All authors Tianqi Zhu, Jun Yang, Jianliang Mao, Cong Li, Shihua Li

Probability

Stochastic heat equations with non-Lipschitz coefficients driven by symmetric $\alpha$-stable Lévy white noise

Juan J. Jiménez

We study the stochastic heat equation on a bounded interval with zero Dirichlet boundary conditions, driven by symmetric $\alpha$-stable space–time Lévy white noise, where $1<\alpha<2$. Both the drift and noise coefficients are continuous and have at most linear growth. For bounded deterministic initial data and every $r>1/2$, we construct a weak solution with càdlàg paths in the negative Sobolev space $\mathbb{H}_{-r}$. We also construct a predictable mild solution on the same stochastic basis and with the same noise. At each fixed deterministic time, this mild solution is a density of the distribution-valued weak solution.
Read the full abstract

Analysis of PDEs

Global strong solutions to the 1D full compressible Navier–Stokes equations with temperature-dependent coefficients and vacuum

Yachun Li, Peng Lu and Zhaoyang Shang

We study the initial–boundary value problem for the one-dimensional full compressible Navier–Stokes system whose viscosity and heat conductivity are given by $\mu(\theta)=\mu\theta^\alpha$ and $\kappa(\theta)=\kappa\theta^\beta$, respectively. The initial density is nonnegative and is allowed to vanish. For $\alpha>2$ and $\beta>2\alpha+4$, we prove the existence of a global strong solution for arbitrarily large compatible $H^2$ data, provided that the prescribed boundary temperature is sufficiently high relative to their size. No velocity or temperature perturbation is required to shrink as the boundary temperature tends to infinity. A shifted initial-layer weight and an exact symmetrization of the two material-derivative equations separate the fast thermal mode from the momentum mode. The resulting estimates are uniform with respect to the positive lower bound used in the approximation and therefore survive the limit to genuine vacuum. The proof further combines a relative-entropy identity centered at the conserved mean density, the physical effective viscous flux, a Zlotnik-type estimate along particle trajectories, and finite-time transport estimates for the density derivatives. The density remains uniformly bounded and the temperature stays uniformly comparable with the prescribed boundary temperature. After the initial layer we establish vacuum-compatible exponential relaxation: the density converges to its conserved mean in every finite $L^p$, the velocity and temperature perturbation decay in $H^1$, and the weighted material-derivative energy decays at rate $c\bar\theta^{1-\alpha}$.
Read the full abstract

Machine Learning

SFlexRCA: Lightweight, Scalable, and Flexible Root Cause Analysis for IIoT Edge Systems

Amr M. Zaki, Farhoud Jafari, Honggeun Ji et al. (+2 more)

Industrial Internet of Things (IIoT) systems generate high-dimensional sensor telemetry from interconnected components, where faults can propagate across the system. To address these challenges, we propose SFlexRCA (Scalable and Flexible Root Cause Analysis), a topology-free RCA framework designed for resource-constrained IIoT environments. SFlexRCA transforms multivariate telemetry into compact orthogonal representations and applies shared lightweight linear modeling, avoiding explicit graph construction, message passing, and per-variable or lag-specific parameter growth. We evaluate SFlexRCA on three publicly available IIoT datasets, BATADAL, SWaT, and WADI, spanning different numbers of monitored variables, temporal characteristics, and training-data regimes. SFlexRCA is compared with 10 statistical, causal, and non-causal baselines in terms of RCA accuracy, training efficiency, inference latency, and memory consumption. In addition, inference efficiency and memory consumption are evaluated on Raspberry Pi 3 and Raspberry Pi 5, while energy consumption is additionally measured on Raspberry Pi 5. We further investigate temporalcontext sensitivity, architectural and loss components, and alternative representations. Notably, SFlexRCA maintains strong localization performance on BATADAL despite its limited normal-operation training data, while its compact shared architecture avoids the parameter growth associated with causal and graph-based approaches. Its lightweight shared architecture further enables efficient deployment on resource-constrained IIoT edge devices. The SFlexRCA code is available at https://github.com/theamrzaki/RootCause- Analysis-Correlation-Attentive-Modeling.
Read the full abstract

All authors Amr M. Zaki, Farhoud Jafari, Honggeun Ji, Komal Sarda, Marin Litoiu

Materials Science

Identification and Structural Decomposition of Hidden Defect Configurations: A Case Study of Charged Oxygen Divacancies in HfO$_2$

Hyunjin Lee, Yoon Kyeung Lee, Keun Heo et al. (+2 more)

Extensive configurational searches have revealed hidden defect structures in crystalline compounds, yet how their local bonding arrangements modify the surrounding atomic environment remains less well understood. We examine this question in hafnium oxide (HfO$_2$), which hosts energetically competing monoclinic (M), tetragonal (T), and polar orthorhombic (PO) polymorphs. For the $c$-axis oxygen-divacancy complex ($T_c$), a local-distortion search with density-functional theory (DFT) relaxation under tetragonal-lattice confinement identifies a hidden family of low-energy configurations ($T_c^{*}$). For the fully charged complex ($2\mathrm{V}_{\mathrm{O}}^{2+}$), structural decomposition using Smooth Overlap of Atomic Positions (SOAP) descriptors reveals mixed M-, T-, and PO-like character, supported by independent analyses of Hf coordination and bond geometry. The idealized $T_c$ reference connects barrierlessly to the lowest-energy mixed-motif $T_c^{*}$ configuration, lowering the energy by $18.4$ meV/f.u. within the confined landscape. After releasing the lattice constraint, we compare transformation pathways to M and PO using solid-state nudged elastic band calculations with cell-relaxed endpoints. The respective barriers increase from $6.9$ and $10.1$ meV/f.u. for the idealized reference to $26.9$ and $15.3$ meV/f.u. for the $T_c^{*}$-derived structure. The larger increase for M makes PO, rather than M, the lower-barrier branch. This case study shows how structural decomposition connects hidden defect configurations with changes in the surrounding atomic environment and kinetic branching between competing polymorphs.
Read the full abstract

All authors Hyunjin Lee, Yoon Kyeung Lee, Keun Heo, Shi Liu, Taehun Lee

Combinatorics

A note on the injective edge coloring of graphs with bounded maximum degree

Lin Tian and Runze Wang

Let $G$ be a graph. Under an injective edge coloring of $G$, any two edges at distance $2$ or belonging to a common triangle receive distinct colors. The injective chromatic index of $G$, denoted by $\chi'_{inj}(G)$, is the smallest integer $k$ such that $G$ admits an injective edge coloring with $k$ colors. Let $\Delta$ be the maximum degree of $G$. Ferdjallah et al.~proved that $\chi'_{inj}(G) \le 2(\Delta-1)^2$. In this paper, we improve this bound by proving that $\chi'_{inj}(G) \le 2(\Delta - 1)^2 -\Delta + 3$ if $\Delta \ge 4$.
Read the full abstract

Combinatorics

Even cycle decomposition thresholds for dense multipartite graphs

Weiyi Sun, Tao Feng and Shikang Yu

Let $r\geq 2$ be an integer. An $r$-partite graph $\Gamma$ with vertex partition $V_1,\ldots,V_r$ is $2$-balanced if there exists a positive integer $n$ such that $n\leq |V_i|\leq 2n$ for every $i\in[r]$. For an integer $\ell\geq3$, let $C_\ell$ denote the cycle of length $\ell$, and define $\hat{\delta}(\Gamma)=\min\{d_\Gamma(v,V_i)/|V_i|:i\in[r],\ v\in V(\Gamma)\setminus V_i\}$. Let $\hat\delta^r_{C_\ell}$ denote the $C_\ell$-decomposition threshold for $2$-balanced $r$-partite graphs, that is, the least nonnegative real number $\delta$ such that, for every $\varepsilon>0$, there exists $n_0$ such that every $C_\ell$-divisible $2$-balanced $r$-partite graph $G$ with $\min_i|V_i|>n_0$ and $\hat\delta(G)\geq\delta+\varepsilon$ admits a $C_\ell$-decomposition. We prove that $\hat\delta^{r}_{C_4}=\frac{2}{3}$ and $\hat\delta^{r}_{C_{2k}}=\frac{1}{2}$ for every $r\geq2$ and every $k\geq3$.
Read the full abstract

Artificial Intelligence

ForkPilot: Self-Evolving Policy for Retrospective Search in Long-Horizon Agents

Xinyue Zeng, Shivam Shandilya, Guilherme Potje et al. (+6 more)

Interactive language-model agents increasingly solve complex tasks through long-horizon, multi-call reasoning, where errors in beliefs or actions can compound across tool interactions. Retrospective search can recover from such failures but is prone to misallocation. Delayed outcomes obscure the contribution of intermediate search decisions, leading to Attribution Complexity, while evolving execution evidence leads to Adaptation Complexity, where previously learned estimates become stale. To address these challenges, we first introduce Search Value Dynamics (SVD), which characterizes the evolving trade-off between the gain and cost of retrospective search. Building on SVD, we propose ForkPilot, a self-evolving two-stage policy-learning framework. In the first stage, ForkPilot learns a search-value policy offline from completed trajectories through automatically constructed outcome comparisons. In the second stage, it makes search decisions based on current observations and then self-evolves by incorporating newly completed trajectories into subsequent policy updates. We evaluate ForkPilot across 6 diverse benchmarks and 7 widely used LLM backbone families, including four open-source families, GPT-5.6 Sol, and Opus 4.8 in a production agentic system, against 9 competitive baselines, including a real-world harness deployment used by hundreds of thousands of paid users. ForkPilot achieves comparable state-of-the-art performance while reducing token usage by up to 59.2%, demonstrating its efficacy.
Read the full abstract

All authors Xinyue Zeng, Shivam Shandilya, Guilherme Potje, Leonardo Nunes, Rakshanda Agarwal, Ranveer Chandra, Emre Kiciman, Dawei Zhou, Tusher Chakraborty

Computation and Language

No Hindsight for LLM Fact-Checkers: Measuring Leakage Channels in Misinformation Detection

Kuan-Hua Wu Lu and Yohanes Andre Setiawan

As automated fact-checking scales on social media, large language model (LLM) verdict scores can look stronger than warranted. One reason is that evaluations mix in information that was not knowable at claim time. Two channels are easy to conflate: outcomes memorized in pre-training and retrieved evidence published after the claim. Yet standard benchmarks rarely separate the two. In this study we measure both channels on AVeriTeC and QuanTemp++ by reconstructing point-in-time evidence conditions and probing for outcome information encoded in model representations. We find substantial evidence of parametric leakage, that can be hidden by the aggregate accuracy, while a simple representation bottleneck reduces this future leakage more efficiently than a mutual-information-based training penalty. We also find that allowing post-claim evidence inflates zero-shot accuracy by 6.3 points in AVeriTeC while the effect is negligible in QuanTemp++, where retrieval provides little post-claim evidence. These results show that misinformation benchmarks can overstate fact-checking performance when they do not account for what information was actually available at claim time.
Read the full abstract

Sound

TS-SP: Learning Speaker-Preserving Representations in Audio Large Language Models

Junjie Li, Zheng Liang, Zhe Li et al. (+2 more)

Audio large language models (ALLMs) can understand speech content, yet their ability to use speaker identity for verification remains limited. We propose TS-SP (Two-Stage Speaker Preservation), a parameter-efficient framework for learning speaker-preserving representations and making them accessible to an ALLM's language-model component. We instantiate and evaluate TS-SP on Qwen2.5-Omni-7B. First, we adapt the audio encoder with speaker identity supervision. We then freeze the adapted encoder and train the language model to compare speakers. Both stages use low-rank adaptation (LoRA), keeping the pretrained base weights fixed. On Vox1-O, TS-SP reduces the equal error rate (EER) from 7.01% for the Paired Loss Adaptation Baseline to 4.37%. EER remains within 4.31–4.79% under unseen prompts. Cross-domain evaluation on CN-Celeb yields a similar EER to the baseline, but lower accuracy at the native decision threshold. These findings support two-stage adaptation for improving speaker verification on the evaluated backbone.
Read the full abstract

All authors Junjie Li, Zheng Liang, Zhe Li, Tianchi Liu, Kong Aik Lee

Nuclear Theory

Nuclear mass table in deformed relativistic Hartree-Bogoliubov theory in continuum, III: nuclei with $8 \leq Z \leq 120$

DRHBc Mass Table Collaboration: Peng Guo, Xiaojie Cao, Kangmin Chen et al. (+89 more)

The mass table in the deformed relativistic Hartree-Bogoliubov theory in continuum (DRHBc) with the PC-PK1 density functional has been established for nuclei with $8 \leq Z \leq 120$, extended from the previous works for even-even nuclei [Zhang et al. (DRHBc mass table collaboration), At. Data Nucl. Data Tables 144, 101488 (2022)] and for even-$Z$ nuclei [Guo et al. (DRHBc mass table collaboration), At. Data Nucl. Data Tables 158, 101661 (2024)]. The calculated binding energies, two- and one-nucleon separation energies, root-mean-square (rms) radii of neutron, proton, matter, and charge distributions, quadrupole deformations, neutron and proton Fermi surfaces, and the blocked neutron (proton) orbitals of odd-$N$ ($Z$) nuclei are tabulated and compared with the available experimental data. A total of 9495 nuclei are predicted to be bound, with an rms deviation of 1.444 MeV from the 2380 mass data. Good agreement with the available experimental pairing gaps, $\alpha$ decay energies, and charge radii is also achieved. The accuracies of the calculated nuclear masses and nucleon separation energies as well as the prediction for drip lines are compared with those obtained by other relativistic and nonrelativistic density functional calculations. It turns out that the DRHBc theory with PC-PK1 provides one of the best microscopic descriptions for nuclear masses. The systematics of nucleon separation energies, pairing gaps, pairing energies, two-nucleon gaps, $\alpha$ decay energies, rms radii, quadrupole deformations, potential energy curves, neutron density distributions, and neutron mean-field potentials are discussed.
Read the full abstract

All authors DRHBc Mass Table Collaboration: Peng Guo, Xiaojie Cao, Kangmin Chen, Qibo Chen, Myung-Ki Cheoun, Yongbeom Choi, Wenmin Deng, Jianmin Dong, Pengxiang Du, Xiaokai Du, Kangda Duan, Xiaohua Fan, Wei Gao, Lisheng Geng, Xi Guo, Yixin Guo, Eunja Ha, Xiao-Tao He, Jinniu Hu, Rongyan Hu, Jingke Huang, Kun Huang, Yanan Huang, Zidan Huang, Xiaofei Jiang, Seonghyun Kim, Youngman Kim, Chang-Hwan Lee, Jenny Lee, Jian Li, Minglong Li, Zhipan Li, Zhengzheng Li, Zhanjiang Lian, Haozhao Liang, Boran Liu, Lang Liu, Xiao Lu, Chenjun Lv, Jie Meng, Myeong-Hwan Mun, Yifei Niu, Zhongming Niu, Cong Pan, Panagiota Papakonstantinou, Jing Peng, Xiaoying Qu, Tianshuai Shang, Xinle Shang, Caiwan Shen, Guofang Shen, Ting Sun, Tingting Sun, Wei Sun, Xiang-Xiang Sun, Chen Wang, Peng Wang, Sibo Wang, Yiran Wang, Yuanyuan Wang, Yumeng Wang, Jiawei Wu, Liang Wu, Xinhui Wu, Xuewei Xia, Yingfeng Xu, Yonghe Xue, Jiangming Yao, Yaochen Yang, Guixiang Ye, Lian Ye, To Chung Yiu, Jianghan Yu, Yangyang Yu, Wentao Zeng, Kaiyuan Zhang, Jinzhe, Zhang, Ning Zhang, Shijie Zhang, Shuangquan Zhang, Wei Zhang, Xiaoyan Zhang, Yanxin Zhang, Ying Zhang, Yingxun Zhang, Zhenhua Zhang, Qiang Zhao, Ruyou Zheng, Chang Zhou, Shan-Gui Zhou, Lianjian Zou

Signal Processing

Beamforming and Trajectory Optimization for Aerial RIS-Aided Secure ISAC via Model Predictive Control

Zhendong Li, Zesheng Zhang, Zhou Su et al. (+3 more)

This paper proposes a novel aerial reconfigurable intelligent surface (RIS)-aided secure integrated sensing and communication (ISAC) architecture. An unmanned aerial vehicle (UAV)-mounted RIS establishes virtual line-of-sight links to bypass blockages, serving legitimate users while sensing potential eavesdroppers. To ensure communication security and sensing reliability, we formulate an infinite-horizon dynamic control problem maximizing the long-term average system secrecy rate and minimizing UAV flight energy. This formulation jointly optimizes the UAV trajectory, active beamforming with artificial noise (AN) at the base station, and RIS phase shifts, subject to strict radar signal-to-noise ratio and UAV kinematic constraints. Due to the highly coupled non-convex variables, solving this problem directly is intractable. Therefore, we propose an online joint optimization framework based on model predictive control (MPC). Within each receding horizon, alternating optimization is employed to decouple the problem into three subproblems, efficiently solved via successive convex approximation and difference of convex programming. Extensive simulations demonstrate that our MPC-based algorithm significantly outperforms static baselines in secrecy rate, exhibiting superior online correction capability and robustness against environmental perturbations.
Read the full abstract

All authors Zhendong Li, Zesheng Zhang, Zhou Su, Ziwei Liu, Ying Wang, Wen Chen

Analysis of PDEs

Hardy Space and Exponential Decay of Inhomogeneous Navier-Stokes System in the 2D Global domain

Xinyu Fan, Boqiang Lv and Xue Wang

We investigate the inhomogeneous incompressible Navier-Stokes system in the 2D global domain with zero initial total momentum, allowing vacuum and densities with finite moments but not necessarily compact support. Exploiting the cancellation structure inherent in the 2D equations along with spatial moment bounds on the density, we derive that the inertial force belongs to the Hardy space and establish weighted Stokes estimates through endpoint Hardy-Littlewood-Sobolev inequalities. These estimates identify the far-field velocity and control the remainder, leading to uniform exponential decay of the velocity throughout global domains. Combined with higher order regularity bounds, we establish the global well-posedness of strong solutions for the density-dependent viscosity case, and prove the uniform regularity of flow maps for measurable density patches in the constant viscosity model.
Read the full abstract