Peer Revenue Past citations are no guarantee of future returns

Papers

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

Computer Science

cs Watch Computer Science (sign in to use)

Tuesday 18 August

Computer Vision

NeuroPath: Brain-Inspired Dual-Pathway Graph Convolutional Networks for Skeleton-Based Action Recognition

Skeleton-based action recognition aims to recognize human actions from sequences of human joint coordinates. Most existing Spatial-Temporal Graph Convolutional Networks (STGCNs) have achieved promising results by modeling skeletal structures with implicit spatial-temporal representations. However, our empirical study reveals a clear performance imbalance across different skeletal modalities, indicating that implicitly coupling spatial and temporal information limits the full exploitation of complementary structural and motion cues. Inspired by the ventral and dorsal pathways in human perception, we propose Dual-Pathway Graph Convolutional Networks (NeuroPath), which adopt a dual-pathway architecture for separate yet collaborative modeling of spatial and temporal information. Specifically, transformation units first convert the input into pathway-specific skeletal representations, allowing each pathway to focus on complementary aspects of human motion. To further capture coordinated joint behaviors and their interrelationships, we introduce a group graph convolution block that dynamically identifies key body parts and models their spatial-temporal dependencies. In addition, inter-pathway dynamic fusion modules integrate complementary inter-modal information across pathways, facilitating higher-level semantic interpretation of actions. Extensive experiments on Kinetics Skeleton 400, NTU RGB+D 60, and NTU RGB+D 120 demonstrate consistent performance improvements, validating the effectiveness of dual-pathway spatial-temporal modeling for skeleton-based action recognition.
Read the full abstract

All authors Kanglei Zhou, Ruizhi Cai, Hubert P. H. Shum, Frederick W. B. Li, Xiaohui Liang

Cryptography and Security

KeyPooling: Measuring Where LLM API Relay Paths Collapse Prompt Cache Isolation

Large language model (LLM) API relays authenticate customers separately but often forward requests through shared provider credentials. Providers scope prompt caches to upstream principals and namespaces, so relay customers mapped to one cache identity can observe each other's cache state. Prior work showed cache sharing at selected endpoints but did not identify which credential, pool, adapter, or nested hop controls the finalidentity. We present KeyPooling, a measurement method that traces customer identity through cache lookup and write, verifies runtime transformations, and tests one predicted identity component at a time. Across five open-source gateways connected to OpenAI and Anthropic, none bound customers to upstream credentials by default; under a shared credential, all five exposed cross-customer cache reads for both providers. Principal and namespace splits, pool associations, and adapter and nested-relay contrasts localized the controlling transformations. In an outcome-independent weekly OpenRouter frame, tests covered 80.5% of eligible token volume and found cross-account reads for 12 of 28 labels carrying 33.7% of volume. On one production route, a controlled procedure recovered eight consecutive target positions without target access. Broader tests identify cache granularity, routing, rate limits, attribution, and budget as conditions for token-by-token recovery, not security controls. We derive a defense contract: every customer must enter a provider-enforced domain, or a namespace derived from authenticated identity must survive every final cache lookup and write. Placing this split after reusable public prefixes preserved most modeled reuse at a 1.7-2.5% cost increase.
Read the full abstract

All authors Bowen Sun, Yixi Cai, Xiaogeng Liu, Zhengyue Zhao, Yinzhi Cao, Chaowei Xiao

Robotics

Reuse Before You Retrieve: Diagnosing Headroom and Complementarity for Test-Time Augmentation of Embodied Multimodal Policies

Frozen vision-language-action (VLA) policies are increasingly improved at test time by sampling additional policy behaviors or introducing external demonstrations. Yet there is little guidance for deciding which intervention a deployed policy actually needs. Additional sampling is useful only when better behavior already exists within the policy's stochastic rollouts and can be identified, whereas retrieval is most useful when the relevant action prior is not reliably represented by the policy. We study this decision through two measurable factors, recoverable headroom and retrieval complementarity, which characterize how much useful behavior is already available to recover and whether an external action prior fills a measurable gap. We evaluate an episode-level retry selector under retryable or parallel execution, together with retrieval across multiple frozen VLA policies and environments. The selector consistently recovers substantial latent capability across all tested VLA backbones on LIBERO, with gains of up to 21.0 success-rate points that closely track recoverable headroom. It also transfers to a different robot and simulator and remains effective under degraded observations, while experiments with autoregressive OpenVLA illustrate the distinction between available headroom and the ability to rank candidate rollouts. Retrieval behaves differently, improving the policy with the largest measured action-prior gap and providing further gains when combined with selection. Together, these results provide an empirical basis for characterizing test-time augmentation opportunities by separating capability that can be recovered from the frozen policy from behavioral priors that may need to be introduced externally.
Read the full abstract

Computer Vision

S$^3$AM: A Single-Stream SAM with Reliability-Calibrated Frequency Adapter for Multi-modal Salient Object Detection

Vision foundation models have recently advanced multi-modal salient object detection (MSOD) through parameter-efficient tuning and prompt learning. However, existing Segment Anything Model (SAM)-adapted MSOD methods often rely on dual-stream encoders or auxiliary prompt generators, leading to redundant computation. Although a single-stream alternative can reduce this cost, early fusion may also propagate noisy or misaligned auxiliary high-frequency cues through the backbone. In this paper, we propose a novel single-stream framework that integrates reliability-calibrated frequency adaptation into the adopted SAM backbone for MSOD. It avoids duplicated foundation backbones while explicitly controlling auxiliary frequency injection. Specifically, we design a mixture of frequency experts module, which uses the stationary wavelet transform to decompose each modality and aggregate cross-modal frequency information. We further introduce a reliability-calibrated frequency adapter with a dual-gate calibration mechanism, which selectively propagates the calibrated residual across transformer stages while jointly controlling its injection strength and cross-modal reliability. A hypernetwork-guided semantic-structural decoder then combines semantic mask features from the adopted backbone with Mamba-based structural detail recovery. Comprehensive experiments on RGB-D, RGB-T, and RGB-NIR salient object detection benchmarks validate that the proposed framework achieves competitive performance with only 12.20M trainable parameters, accounting for 5.4% of the total parameters. The code will be available at https://github.com/xuboyue1999/SSSAM.
Read the full abstract

All authors Ruichao Hou, Boyue Xu, Tongwei Ren, Dongming Zhou, Gangshan Wu, Jinde Cao

Artificial Intelligence

When AI Designs AI: Innovation or Imitation?

Recent advances in LLM agents have made them increasingly capable of designing methods for complex AI tasks. This raises two central questions about agent-designed methods relative to human-designed methods: how well they perform, and how different their algorithmic designs are. To study these questions, this paper introduces an analysis that derives task-specific algorithmic design spaces from human-designed methods, maps both human- and agent-designed methods into these spaces, and quantifies their algorithmic differences at the module level. Widely used LLM agents are evaluated on a suite of representative, open-ended AI tasks spanning multiple modalities, and the methods they design are analyzed in terms of both task performance and algorithmic differences from human-designed methods. Experimental results show that current agents can occasionally match or surpass human state-of-the-art (SOTA) performance (10/72 configurations), but such success does not generalize reliably across tasks or agents. Moreover, 96.8% of agent-designed methods fall within human-derived algorithmic design spaces, largely recombining algorithmic choices found in human-designed methods, while nearly half exactly match an existing human algorithmic design. Taken together, these findings suggest that although current agents can occasionally match or surpass human SOTA performance, their algorithmic designs remain within human-derived algorithmic design spaces, reflecting the reuse and recombination of algorithmic choices.
Read the full abstract

All authors Yikang Yang, Zhengxin Yang, Luzhou Peng, Minghao Luo, Yanqi Kan, Wanling Gao, Jianfeng Zhan

Emerging Technologies

Adaptive Participation Under Statically Equivalent Incentives in Distributed Demand Response Systems

Aggregators recruit distributed energy resources with settlement rules and participation payments. Such designs are normally validated at fixed points: zero participation must cease to be an equilibrium, and truthful capability reporting must remain a best reply. We ask whether those checks determine the participation that owners reach once they adapt from the settlements they receive. In a five-unit event with fixed dispatch, payment rule and penalty, we vary only how a scarcity-contingent participation payment decays with the capability others have declared. Of two decay structures that agree on all five static criteria, one reaches full participation from a collapse initialization in 96 of 96 seeds and the other in none, within an 8000-round horizon and with disjoint 95% confidence intervals. The difference lies in the payoffs offered at partial participation, which the static criteria never evaluate; it is a property of experience-based feedback and closes when counterfactual payoffs are supplied. Because those payoffs make each unit's settlement depend on what the others declared, we also ask what survives when the mechanism is distributed. Running the aggregator and the five units as separate processes reproduced the centralized reference at every round, and a deliberately misattributed declaration was detected although every message was delivered.
Read the full abstract

All authors Xun Shao, Ryoichi Inoue, Shinken Takekawa, Go Hasegawa

Artificial Intelligence

SAGE: Self-Evolving Storyboard Skills via Attribution-Guided Rule Evolution

Storyboards turn screenplays into visual shot plans for automated short drama production. Professional storyboarding relies on tacit directorial expertise and remains an industrial bottleneck. Large language models can automate this step, but methods for supplying directing knowledge face three challenges: (1) Knowledge acquisition: the craft remains implicit in exemplars or must be written manually. (2) Knowledge refinement: authored knowledge is not evaluated against execution outcomes, and opaque generation prevents feedback attribution to the knowledge behind each decision. (3) Knowledge injection: injecting all knowledge exceeds usable context, while manual selection for every narrative group does not scale. We present SAGE (Skill with Attribution-Guided Evolution), a deployed framework that learns, attributes, evolves, and routes directing knowledge from expert demonstrations. SAGE derives rules that are independent of episode content by contrasting each training screenplay with its expert storyboard. During generation, the model records each narrative group's adopted rules. Combining these records with localized feedback enables targeted updates to individual rules. Evolved rules form scenario packages with a routing index, so each group retrieves only a bounded set appropriate to its situation without expert intervention. On 18 test episodes across three genres, SAGE scored 77.8 on a rubric validated by experts, versus 77.1 for professional directors. Deployed for 14 days on Virtual Film Studio, SAGE produced 1,344 narrative group outputs; 87.2 percent were accepted without substantive edits, and the production team recorded over 83 percent less authoring time per episode. We release PROSE, the first public dataset pairing screenplays with storyboards by professional directors across 68 episodes: https://github.com/creDreams/PROSE.
Read the full abstract

All authors Maolin Ran, Xiaoyang Lu, Jiaqi Liu, Jian Wang, Weiwen Liu, Jianghao Lin, Yong Yu, Weinan Zhang

Networking

LoRIS: LoRaWAN-based IoT Platform for Sustainability Monitoring in Hotels

The hospitality sector is a major source of global greenhouse gas emissions, water stress, and waste generation, yet sustainability reporting in hotels remains constrained by coarse, manually collected operational data. We present LoRIS (LoRaWAN-based IoT platform for sustainability monitoring in hotels), a LoRaWAN-based sensing system that delivers high-resolution measurements of resource consumption, environmental conditions, and guest behaviour across geographically distributed hotel properties. The architecture follows the canonical LoRaWAN reference model and is built for the operational realities of hospitality deployments: restrictive hotel IT policies, guest privacy expectations, rapid and reversible installation, and multi-year battery operation. Privacy-by-design guides modality selection and deployment zoning, and end-to-end encryption protects data from sensor to dashboard. This system has been running since February 2022 and currently spans 850 sensors of 19 types across 21 sites in Australia and Slovenia, covering both the AU915 and EU868 regulatory regions. The platform has generated over 202 million sensor records and ingests approximately 245,000 uplink messages per day on managed serverless infrastructure. Our system has been successfully used for seven field studies spanning food waste, energy consumption, and water consumption, including controlled intervention experiments that measure environmental outcomes and guest satisfaction in parallel. This system shows that LoRaWAN sensing can be deployed at scale in operational hotels without compromising guest experience or privacy.
Read the full abstract

All authors Yash Pandey, Angus Gray, Reza Serati, Oscar Zhu, Emil Juvan, Anna Zinn, Danyelle Greene, Qingqing Chen, Sarah MacInnes, Siamak Layeghy, Sara Dolnicar, Marius Portmann

Machine Learning · Cross-listed from stat.ML

Online Generalized Sparse Regression: How Does Overparametrization Help?

Regularized sparse regression has been extensively studied in the offline setting, but online formulation remains relatively under-explored. This gap stems from four key challenges: (i) the infeasibility of dynamically updating the regularization parameter in every online round, (ii) managing storage and memory complexity, (iii) enabling real-time computation via closed-form updates rather than solving full optimization problems at each round, and (iv) achieving optimal statistical guarantees under realistic assumptions. In this paper, we propose an online generalized-sparsity-constrained regression framework, focusing on online cardinality-constrained linear regression and low-rank matrix sensing. Unlike online regularized regression, our constrained formulation eliminates the need for dynamic parameter tuning. We introduce an efficient online hard-thresholding algorithm that performs closed-form updates and requires storing only summary statistics, making it computationally, memory, and storage efficient. Despite the inherent nonconvexity and combinatorial nature of the formulation, our algorithm achieves global convergence at the optimal statistical rate under realistic assumptions, provided that the projection set is properly overparameterized. Numerical experiments demonstrate that our method consistently outperforms state-of-the-art alternatives.
Read the full abstract

Systems and Control

Infinite-Horizon Sparse Optimal Control: Solution through a Finite-Horizon Subproblem and Its Receding-Horizon Implementation

Sparse optimal control is considered in the infinite horizon. In the literature, sparse control has been considered mostly in a finite horizon for its formulation into a finite-dimensional optimization problem. It is shown in this paper that an optimal solution of the infinite-horizon sparse control problem can be obtained through a solution of some finite-horizon subproblem. This is due to sparsity of the optimal solution in the sense that the optimal control input is constantly equal to zero at its tail. An estimate is given on the horizon length required by this subproblem and its adaptive choice is also discussed. Implementation with a receding-horizon technique is considered and its optimality and sparsity are guaranteed.
Read the full abstract

Information Theory

Dynamic Question Design for Efficient Estimation of Aggregate Human Preferences

This paper addresses the problem of efficiently estimating aggregate human preferences by dynamically adapting questions based on respondents'answers. To this end, we formulate and address two sub-problems: preference estimation and question design. First, regarding preference estimation, we model respondents' preferences and estimate them using Bayesian estimation, employing a particle filter as a computationally efficient approximation. The main theoretical contribution to this sub-problem is to analyze the preference estimation error using an information-theoretic approach, deriving a theoretical lower bound for the error. Second, regarding question design, we formulate the design problem as an Expected Information Gain maximization problem and employ an epsilon-greedy strategy to solve the problem in a computationally efficient way. We theoretically analyze the search efficiency of the approach, demonstrating that it achieves higher efficiency than a random search. Finally, we verify the effectiveness of the proposed method through numerical simulations.
Read the full abstract

Robotics · Cross-listed from physics.flu-dyn

Optimal control of a swimming robot based on Purcell's microswimmer model

Purcell's swimmer is a well-known planar model of a swimming microorganism, governed by low Reynolds number hydrodynamics, which is comprised of three rigid links connected by actuated rotary joints. This model has been analyzed as a robotic locomotion system governed by first-order nonlinear dynamics with a periodic input (gait) of the two joint angles. In this work, we present a robotic macro-scale realization of this three-link swimmer moving in a highly viscous fluid. We propose a simple variant of Purcell's theoretical model with non-slender links and a central rigid sphere which represents the added drag of the robot's central flotation block, and calibrate the model's parameters to fit experimental measurements. Next, we apply optimal control formulation based on Pontryagin's Maximum Principle (PMP) in order to find optimal gaits that maximize the displacement per cycle under bounds on the joint angles. Employing a differential geometric method that transforms the problem to area integral enclosed by the gait trajectory in the plane of joint angles, enables visual interpretation which explains topological changes in displacement-optimal gaits upon varying the bound on the joint angles. We then apply PMP formulation to the problem of maximizing Lighthill's energy efficiency in order to obtain a boundary value problem (BVP) whose solution gives efficiency-optimal gaits for Purcell's swimmer model, as well as its variant with a central sphere. Finally, we utilize numerical methods such as parameterizing the input gait as a truncated Fourier series, as well as GPOPS-II solver, to produce sufficient initial guess values for solving the BVPs and obtaining efficiency-optimal gaits.
Read the full abstract

Computation and Language

From Entity Mentions to Tone: An LLM-Based Pipeline for Media Bias Analysis

This paper presents a pipeline for analyzing media bias and framing in online news. The pipeline groups articles into topics and events, adds named-entity and sentiment annotations, and compares news sources through people mentions, source-level tone, and event-level coverage patterns. We apply it to 8,358 Albanian news articles collected from GDELT and compare the resulting annotations with GDELT's automated annotations. The results show moderate agreement for sentiment and entity extraction, as well as additional person-entity pairs that can potentially support the bias analysis. We compare two annotation prompts and find that stricter sentiment-validation rules remove label-score inconsistencies but increase execution time and reduce annotation coverage. Based on these results, the simpler prompt is used for the rest of the analysis. We have provided sample analysis on source-level framing pro les, person-level tone differences across sources, and event-level gatekeeping and coverage indicators. These outputs show how the same news collection can be used to examine what sources cover, how they describe public figures, and where coverage is concentrated. The approach is particularly useful in settings where manually verified datasets or specialized language tools are limited.
Read the full abstract

Robotics

EATR-Stereo: Embodiment-Aware Token Routing of Paired Stereo Evidence for Humanoid Vision-Language-Action Control

Long-horizon humanoid vision–language–action (VLA) control with head-mounted stereo cameras requires visual interfaces that can exploit complementary views while maintaining compatibility with pretrained representations. Existing interfaces often discard complementary stereo evidence or fuse additional observations without preserving the native primary-view pathway and adapting auxiliary information to robot embodiment. We present EATR-Stereo, an embodiment-aware token-routing framework that retains primary-view tokens and constructs primary-aligned Cross-View Auxiliary Tokens (CVATs) by querying the synchronized auxiliary-view token sequence. A body-segmented proprioceptive encoder further conditions token-wise auxiliary usage on robot configuration history, enabling selective incorporation of stereo evidence during action generation. The routed auxiliary stream augments the language and primary-visual context of a pretrained VLA while keeping its vision–language model frozen. On a 33-DoF physical humanoid with a 37-D proprioceptive state, we evaluate nine configurations in over-100-s search–approach–grasp–place–return tasks. EATR-Stereo achieves 60.0% full-task success, 100.0% grasp success, and 80.0% stage success. Under severe asymmetric occlusion, it improves recovery to 80% compared with 30% for CVAT alone. Ablation studies further show the importance of preserving primary tokens and combining cross-view auxiliary features with structured proprioceptive routing. These results demonstrate that selectively routed paired stereo evidence improves spatial grounding for reliable long-horizon humanoid VLA control.
Read the full abstract

All authors Songwei Wu, Rui Zhao, Fan Yang, Zhongqiang Nie, Zhiduo Jiang, Wandong Sun, Yuwei Li, Jian Hu, Yang Liu, Hong Liu

Machine Learning

Causal Local States: Scalable Simultaneous Causal Network Inference and Forecasting for Dynamical Systems

Machine learning methods predict many real-world systems with remarkable accuracy, but they are typically treated as black boxes that offer no insight into which interactions drive the dynamics. Causal discovery methods reconstruct the interaction network from observational data, but without regard to whether the inferred structure supports prediction. Existing approaches combining both tasks rely on a single global hyperparameter, such as a causal threshold or a fixed neighborhood size, which cannot recover the structure of heterogeneous systems. Here we introduce causal local states (CLS), a framework that simultaneously infers an approximate Granger-causal interaction network and forecasts the system dynamics. For each node independently, we select the smallest set of neighbors that allows a predictive model to forecast the node near-optimally, and the resulting neighborhoods are then combined for a forecast of the full system. On three benchmarks of increasing difficulty, we achieve reconstruction of the underlying networks with high fidelity and forecasts on par with a model that is supplied with the true network, providing a step toward explainable and scalable forecasting of complex systems.
Read the full abstract

All authors Jonas Braun, Fabian Fischbach, Daniel Köglmayr, Sebastian Baur, Christoph Räth

Human-Computer Interaction

From Substitution to Scaffolding: Breaking the Self-Reinforcing Harm Cycle of AI in Education (and Beyond)

Artificial intelligence is being adopted in educational settings faster than its consequences are understood. We argue that the central risk is misalignment: AI that eliminates human effort erodes the very capacities education is meant to build. We organize this risk into an integrative framework of four interrelated dimensions -cognition, agency, emotional well-being, and ethics- linked by a self-reinforcing cycle where cognitive offloading reduces effort, weakens agency, and compounds emotional and ethical harm. We ground the framework in the perspective of a small cohort of students: an exploratory analysis of 49 International Baccalaureate argumentative essays about the impact of AI reveals that learners perceive these risks, with $80\%$ of essays reporting that AI reliance reduces thinking. At the same time, the essays articulate a consistent vision of the AI the students want: systems that support rather than replace learning by withholding immediate answers, prompting recall, and encouraging reflection through questions instead of solutions. These desiderata closely align with established principles from the learning sciences. Building on these insights, we propose a single design principle, scaffold, do not substitute. We argue that this principle extends beyond education. It represents a broader challenge for the AI ecosystem: any system that mediates human thinking can either weaken human capabilities through substitution or strengthen them through scaffolding. We conclude by outlining a research agenda for developing AI systems that foster enduring human capacity, an imperative not only for learners but, ultimately, for democratic societies.
Read the full abstract

All authors Lucile Favero, Juan Antonio Pérez-Ortiz, Tanja Käser, Nuria Oliver

Machine Learning · Cross-listed from stat.ML

Nonlocal Transition Kernel for Efficient Learning of Restricted Boltzmann Machines

Learning restricted Boltzmann machines (RBMs) is computationally challenging because it requires expectations whose exact evaluation is generally intractable. The expectations are typically evaluated using a sampling approximation based on blocked Gibbs sampling (BGS), which is a local Markov chain Monte Carlo transition kernel. However, the locality of BGS can lead to poor sampling quality when the RBM has high energy barriers, thereby degrading learning performance. Deep tempering (DT), which performs parallel tempering over a sequence of learnable RBMs including the training RBM, alleviates this locality issue. However, DT algorithmically requires multiple steps to move through the RBM sequence to achieve a nonlocal transition. In this paper, we propose a transition kernel defined over the RBM sequence used in DT. The proposed kernel has a round-trip structure over the sequence, enabling nonlocal moves within a single transition while leaving the RBM sequence invariant. Numerical experiments show that the proposed kernel performs nonlocal transitions more frequently and achieves higher sampling quality with fewer transitions than BGS and DT. We further verify that learning based on the proposed kernel is more stable and mitigates the training failures observed with BGS- and DT-based learning.
Read the full abstract

Systems and Control · Cross-listed from eess.SP

Multi-Sensor Edge Angle Detection for Performance Analysis in Ski Jumping

In ski jumping, performance during the gliding phase depends on achieving an aerodynamic posture that maximizes the lift-to-drag ratio. In the V-style technique, the ski edge angle is a key determinant. Reducing the edge angle flattens the skis, increases their effective surface area, and improves aerodynamic lift, ultimately contributing to longer flight distances. Ski edge angles are biomechanically constrained by the limited range of ankle inversion. Current sensing solutions widely quantify these angles using multi-system approaches that combine sensor signals through geometric relations. Such configurations require instrumentation on both the boot and the ski, altering mass distribution, affecting balance during flight, and increasing system complexity. To overcome these limitations, this work presents a wearable sensing system that measures both boot inclination and ski edge angle without modifying the ski surface. Two ultrasonic Time of Flight (ToF) sensors and an in-shoe Inertial Measurement Unit (IMU) are integrated into a single boot-mounted unit. Edge angles are estimated by combining ultrasonic distance measurements with IMU data through geometric reconstruction of the boot-ski configuration. Laboratory experiments demonstrate an angle resolution of 0.4500{\deg}, a Mean Absolute Error (MAE) of 0.2640{\deg}, and a coefficient of determination exceeding 99% when compared with reference measurements, indicating strong linear agreement between the two modalities. The system achieves an end-to-end latency of 30.31 ms, enabling real-time feedback suitable for athlete training, while consuming 1.28 mW of power. With a total weight of only 18.6 g the proposed system enables unobtrusive measurement of ski edge angle and boot orientation.
Read the full abstract

All authors Ivan Simeonov, Lukas Schulthess, Hanna Mueller, Marc Nölke, Michele Magno, Luca Benini, Christoph Leitner

Computer Vision

NGS-Marker: Robust Native Watermarking for 3D Gaussian Splatting

With the rapid development and adoption of 3D Gaussian Splatting (3DGS), the need for effective copyright protection has become increasingly critical. Existing watermarking techniques for 3DGS mainly focus on protecting rendered images via pre-trained decoders, leaving the underlying 3D Gaussian primitives vulnerable to misuse. In particular, they are ineffective against Partial Infringement, where an adversary extracts and reuses only a subset of Gaussians. In this paper, we propose NGS-Marker, a novel native watermarking framework for 3DGS. It integrates a jointly trained watermark injector and message decoder, and employs a gradientbased progressive injection strategy to ensure full-scene coverage. This enables robust ownership decoding from any local region. We further extend NGS-Marker with hybrid protection (combining native and indirect watermarks) and support for multimodal watermarking. Extensive experiments demonstrate that NGS-Marker effectively defends against partial infringement while offering practical flexibility for real-world deployment.
Read the full abstract

All authors Hao Qin, Yukai Sun, Luyuan Chen, Mengxu Lu, Feng Zhang, Ming Kong, Zhenhong Du, Qiang Zhu

Cryptography and Security

Decomposition Attacks Across Unlinkable Identities: Limits of Stateful Defenses for LLM Services

Most large language model services use stateless defenses, which judge only the current request, to refuse harmful tasks. Decomposition attacks exploit this limitation by splitting a harmful task into individually permissible requests and combining their answers. Defending against them therefore requires a stateful monitor that considers requests together. If it can group all requests for one attacker task, it can stop the attack. However, attackers can use unlinkable identities and combine answers elsewhere, leaving no reliable grouping signal. We ask whether decomposition attacks can still be stopped under this setting. For a fixed attack strategy without retries, we prove that the achievable security and utility tradeoff depends entirely on how benign requests for the same capabilities are grouped. Persistent, recognizable groups permit a useful defense; fresh, indistinguishable groups do not. When attackers can retry and learn from Allow/Block decisions, this useful operating point disappears: the feedback reveals what passes but not whether a block was correct. Experiments on 91 executable tasks and 11,393 capability-matched benign requests support these results. Under a 1% denial cap for these requests and a 0.5% cap for unrelated background traffic, all ten tested policies, including one privileged policy with an exact request-to-operation map, either fail to stop attacks or exceed the budget. On defense-unseen task families, attack success is at least 99% after one attempt and 100% after two. Effective defenses therefore require additional evidence or mechanisms tied to grouping, such as reliable identity linkage, costs for fresh identities, or control over answer use.
Read the full abstract

All authors Bowen Sun, Zhengyue Zhao, Xiaogeng Liu, Yinzhi Cao, Chaowei Xiao

Artificial Intelligence

Structure-Internalized Rule Language Model for Faithful Knowledge Graph Reasoning

Knowledge Graph Reasoning (KGR) aims to discover latent facts by leveraging the structural evidence available in KGs, posing a challenge to the structural semantic understanding capability of KGR models. Recent studies have demonstrated that Large Language Models (LLMs) can achieve remarkable progress on KGR tasks via flexible in-context learning. However, the inherent representation inconsistency between KG structural context and LLM parametric knowledge remains inadequately addressed. This limitation prevents LLMs from effectively perceiving reasoning evidence that aligns with KG constraints, which undermines both the effectiveness and faithfulness of reasoning. We refer to this problem as reasoning evidence perception drift of LLMs over KGs. To address this problem, we propose a Structure-Internalized Rule Language Model (SIRLM), which centers on structural rule generation to couple the parametric learning of structural knowledge with the faithfulness evaluation of reasoning logic, enabling LLMs to anchor tightly to KG-grounded evidence. Specifically, we first design a Structure-Internalized Rule Generator (SIRG), which incorporates an in-context learning block augmented with a structural relation memory to coordinate structural and parametric knowledge. Furthermore, we equip SIRG with a KG tokenizer based on structural invariance learning and a neuro-symbolic reasoner based on rule-constrained message propagation. These components provide SIRG with learnable structural representations and faithful rule-execution feedback, respectively. Our SIRLM can be seamlessly integrated into standard LLM training paradigms, such as SFT and GRPO. Extensive experiments against 17 state-of-the-art KGR methods on 36 datasets demonstrate the significant superiority of SIRLM.
Read the full abstract

All authors Xingrui Zhuo, Jiapu Wang, Manzong Huang, Gongqing Wu, Xindong Wu

Cryptography and Security

FESC: Remodeling Long-Context Private Inference with Encrypted State-Space Models

Processing long, sensitive documents with machine-learning models requires efficient, privacy-preserving long-context inference. Prior private inference systems optimize or distribute encrypted Transformer attention, but its quadratic token-pair work remains the bottleneck as sequence length grows. Selective state-space models (SSMs) offer linear-time recurrence, yet direct encrypted implementation incurs linear multiplicative depth, sequence-wide state residency, or dense FHE-MPC conversion. We present Factorized Encrypted Scan-Contract (FESC), a hybrid FHE-MPC system for private long-context selective SSM inference. Its factorized scan-contract keeps input-dependent transitions compact across conversion boundaries, composes them without dense expansion, streams state chunks on demand, and contracts outputs before conversion. We demonstrate interface compatibility of the scan-contract implementation across invariant and selective SSM architectures. For our Mamba-2 instantiation, we design GPU-optimized CKKS kernels for linear computations, MPC protocols for SiLU, softplus, exponential, and RMSNorm, with approximation-aware fine-tuning. To our knowledge, FESC is the first private long-document inference system to complete native end-to-end execution at $L \geq 1{,}024$ on a single GPU. At $L = 2{,}048$, a 12-layer Mamba-base model completes inference in 77.3 minutes on one A100 GPU with a peak memory footprint of 32.7 GB, while maintaining near-plaintext accuracy on the evaluated long-document tasks.
Read the full abstract

All authors Yufan Zhu, Chao Jin, Khin Mi Mi Aung, Xiaokui Xiao

Numerical Analysis

Improved Convergence of Multilevel Moving Least-Squares Approximation

Moving least-squares approximation is a popular method for approximating multivariate functions from given discrete data. For higher accuracy higher degree polynomials have to be used, resulting also in higher computational cost and numerical instabilities. Recently, the combination of low-order moving least squares with a multilevel scheme showed superior numerical behavior. In this paper we will prove, amongst other things, that such a combination of moving least-squares with a multilevel scheme indeed leads to improved convergence results, at least if the data sites form a regular grid.
Read the full abstract

Machine Learning

General Semantic Knowledge Infusion for Spatio-Temporal Traffic Forecasting

Although Graph Neural Networks (GNNs) have made significant advances in spatio-temporal traffic forecasting, their performance is limited when they rely solely on sensor proximity or road-network topology. This paper presents a spatio-temporal prediction framework, developed to incorporate knowledge in various forms. This framework aims to improve sensor-level, contextual understanding of the environment. A general-purpose knowledge graph (e.g., Wikidata) is used to create semantic subgraphs around traffic sensors and generate knowledge graph embeddings that capture meaningful relationships, such as nearby points of interest, administrative hierarchies, and the functional roles of locations. These embeddings are then fused with conventional traffic sensor graphs to provide additional adjacency matrices informed by semantics. This allows GNNs to learn the semantic context beyond physical connectivity. This study differs from previous research in two key ways. Firstly, rather than proposing a novel GNN architecture, it demonstrates the general impact of external knowledge on prediction accuracy. Secondly, experiments with well-established traffic forecasting approaches show that external knowledge provides additional information that street network data alone cannot convey. The results show that integrating data from general-purpose knowledge graphs and sensor networks through data fusion can enhance the prediction accuracy of traffic forecasting models, and offers a potential pathway toward improved interpretability.
Read the full abstract

All authors Mattis Thor Straten, Yannick Wolker, Steffen Strohm, Prathvish Mithare, Ralf Krestel, Matthias Renz

Artificial Intelligence

Depth Enables Local Entropy: Quadratic Depth Dependence in Deep Variation-Norm ReLU Regression

We study Gaussian regression over the explicit vector-valued Parhi–Nowak deep-RBV^2 architecture with depth L, width w, layer-sum variation budget A, and output bound B. For this O(L w^2)-parameterized architecture, the known lower and upper bounds differ by one factor of depth. We construct a local packing showing that the quadratic depth dependence is intrinsic under an explicit sample-size-dependent radius condition. The packing has log-cardinality Omega(L^2 w^2 log w); its codewords lie in an O(lambda) L^2 ball and are pairwise Omega(lambda)-separated. The main ingredients are a bias-corrected bounded-coefficient approximation theorem and balanced amplification: multiplying a depth-D ReLU network by q can be implemented using one constant channel so that every coefficient grows by only q^(1/D). Translation to vector-valued RBV^2 blocks then has layer-sum cost O(D w^2 q^(1/D)). Gaussian Fano yields a radius-explicit lower bound governed by the output, testing, and representation scales. Under A=B=R, sigma proportional to R, and the stated radius condition, this gives minimax risk at least of order L^2 w^2 log(w) R^2/n. A pseudodimension-based finite-net upper bound gives O-tilde(L^2 w^2 R^2/n) for unbounded Gaussian responses. Thus the minimax risk has quadratic polynomial dependence on depth, up to logarithmic factors, and exhibits a transition to representation-limited behavior at smaller radius.
Read the full abstract

Artificial Intelligence

Task-Aware Harness Provisioning for LLM Agents in Mission-Critical Infrastructure Operations

LLM agents have been widely adopted to operate mission-critical infrastructure (MCI). These agents normally rely on a harness that determines what information they can access, which tools they can use, and what actions they can take. Existing systems often expose the same comprehensive harness to every task, which may not be necessary and cause resource wastes. In this paper, we focus on the identification of optimal harness configurations, and view it as a resource-matching problem between what each task requires and what the harness provides. To measure this match, we classify MCI tasks based on the mathematical representation of the underlying system and rank harness configurations by the amount and type of information they provide. We then construct task-to-harness mappings from two sources: mining research literature and measuring controlled agent execution. Leveraging the measured mapping, we propose a new harness provisioning algorithm: map-guided escalation. It begins with a task-specific harness and expands to full provision only after a failed self-check. We evaluate our method in two representative MCI tasks: in liquid cooling, it improves the agent accuracy from 0.652 under full provision to 0.715 and achieves accuracy comparable to Reflexion with 48% fewer tokens; In power grids, full provision remains accuracy-optimal, while map-based provisioning offers lower-cost alternatives. These findings show that harness provisioning follows a domain-dependent accuracy-cost Pareto frontier rather than a universal optimum.
Read the full abstract

All authors Liangtao Lin, Qingang Zhang, Zhaomeng Zhu, Tianwei Zhang, Yonggang Wen

Robotics

UniReflex: Plug-and-Play Force Control for Pretrained Generative Policies via Fast-Slow Reflex

Generative imitation learning policies excel at trajectory planning but lack closed-loop force regulation, while directly incorporating force modalities often requires redesigning or retraining the network. We present UniReflex, a universal plug-and-play framework that equips frozen generative policies with variable impedance control (VIC) for contact regulation, guided by force-direction intent collected during demonstration, without further slow-backbone fine-tuning. By non-invasively intercepting deep latent representations from the action head, UniReflex drives a fast reflex network that decouples active force exertion from external interaction response. This scheme predicts normalized anisotropic stiffness directions for directional compliance allocation. Furthermore, UniReflex integrates an adaptive gating mechanism that enables seamless transitions between position-dominant planning and force-dominant execution. Real-world bimanual experiments demonstrate that UniReflex significantly improves contact stability and success rates while preserving original position accuracy. Our approach achieves 25-66x lower per-step backward latency relative to joint training strategies on the evaluated backbones.
Read the full abstract

All authors Yan Huang, Shoujie Li, Ziwu Song, Wenbo Ding

Computational Complexity

A Simple Algebraic Proof of the PCP Theorem

We give the simplest known algebraic proof of the PCP theorem, involving only ingredients like code concatenation, polynomial interpolation, and polynomial multiplication. Specifically, we prove that graph 3-coloring has a polynomial-sized proof that can be verified by a verifier tossing logarithmically many coins and querying a constant number of bits in the proof. In particular, our proof does not involve any PCP compositions; notably, it does not invoke the NP-completeness of any fixed problem, such as SAT or 3-coloring, in the construction of the verifier. The main innovation in our work is a clean, coding theoretic, way to encode univariate polynomials that allows us to implement “low-degree testing” using just a constant number of bits of queries. Insights from recent attempts to simplify the PCP proof by the authors (STOC 2026) and Goldreich (ECCC 2025) allow us to observe that low-degree was the key bottleneck in converting previous algebraic constructions of the PCP verifier into a constant query PCP. Thus, by overcoming this bottleneck, we get the full PCP verifier using elementary and self-contained steps. As concrete support for the claimed simplicity, we include the full pseudocode of the PCP verifier, assuming finite field arithmetic, and a full description of the completeness (aka “honest”) prover, assuming multivariate polynomial arithmetic including interpolation and evaluation, that fit in about a page each.
Read the full abstract

All authors Prashanth Amireddy, Amik Raj Behera, Srikanth Srinivasan, Madhu Sudan, Sophus Valentin Willumsgaard

Computer Vision

Counterfactual Anatomy-guided Spatial-Temporal Decoding for Annotation-Free Hallucination Mitigation in Medical VLMs

Medical vision-language models (Med-VLMs) have demonstrated strong performance on medical visual question answering, yet they remain prone to hallucination, generating clinically unsupported statements that are insufficiently grounded in image evidence. Mitigation methods applied during decoding offer a practical solution, but they typically lack anatomical awareness or rely heavily on ground truth annotations, which limits their applicability. We propose Counterfactual Anatomy-guided Spatial-Temporal decoding (CAST), a framework that operates entirely during inference and requires no manual annotations for anatomically grounded hallucination mitigation. CAST automatically discovers anatomical regions relevant to the given query through broad medical segmentation. It then selects a compact, causally informative area using counterfactual intervention based on the drop in answer likelihood under occlusion. Guided by this chosen region, CAST performs a unified contrastive decoding process, combining classifier-free guidance to correct spatial attention with stepwise temporal contrast to regulate generation dynamics. Experiments on the SLAKE and MIMIC-CXR datasets across three Med-VLMs demonstrate that CAST consistently outperforms strong baselines and surpasses decoding strategies reliant on ground truth. Our results indicate that compact, automatically selected regions provide highly effective contrastive guidance without expert annotations, offering a practical and generalizable solution for improving spatial grounding and reducing hallucinations. Code is available at https://github.com/csyifan/CAST.
Read the full abstract

All authors Yifan Lu, Adinath Dukre, Abhijit Das, Ziyun Zou, Haolin Yang, Yutong Xie, Imran Razzak

Computer Vision

SemComp-Bench: Benchmarking Semantic Task Completion in Video Generation

We introduce Semantic Task Completion Video Generation, an outcome-oriented video generation task. Under this formulation, success requires both achievement of the intended outcome and semantic grounding. Semantic grounding characterizes the correspondence between the reference image and the generated outcome in terms of high-level semantics relevant to the task. Evaluation focuses on the generated outcome and requires neither the presentation of a complete sequence of intermediate task steps nor conventional appearance consistency with the reference image. To support systematic evaluation, we construct SemComp-Data, an evaluation dataset covering six domains. Each instance comprises a reference image, a detailed instruction, a brief instruction, and an outcome-centric video clip. A scalable four-stage curation pipeline converts raw videos into standardized SemComp-Data instances. We further introduce SemComp-Bench, an evaluation protocol that uses a vision-language model (VLM) to answer structured binary questions. SemComp-Bench reports the OA Score and the GR Score for Outcome Achievement and Generation Reliability, respectively. Experiments on representative video generation models show that achieving intended outcomes while maintaining task-relevant semantic grounding in reference images remains challenging.
Read the full abstract

All authors Keyu Tu, Zhuowei Chen, Mengqi Huang, Yuxin Wang, Jiahao Zhu, Zhendong Mao, Yongdong Zhang