Searcharxiv⌕ Search

arXiv subjects

Zhen Xu

Publications and source records attributed to Zhen Xu.

At least 37 records · Page 2Linked to original sources
arXiv2026

Does AI Reviewer See the Full Picture? Attacking and Defending Multimodal Peer Review

The integration of Large Language Models (LLMs) and Multimodal LLMs (MLLMs) into scientific peer-review workflows introduces novel and significant risks for adversarial manipulation, especially given the multimodal nature of scientific papers where figures, not just text, convey core evidence. This creates a significant gap: current robustness studies on AI peer-review are overwhelmingly text-only. Moreover, the problem is distinct from standard jailbreaking, as a peer-review attack seeks to induce a domain-specific, targeted failure (e.g., "inflate this score") rather than a general safety policy violation, for which no practical defenses exist. To address this, we introduce PaperGuard, the first comprehensive benchmark designed to systematically evaluate and defend AI-generated peer-review against these domain-specific, cross-modal attacks. Our framework is built on three pillars: (1) a new multimodal peer-review dataset spanning multiple scientific domains; (2) a unified suite of attacks, including black-box prompt injections and white-box perturbations, specifically designed to target both text (GCG) and figures (PGD); and (3) a practical defense, motivated by the long-context challenge of academic papers, that uses chunk-based embedding search to efficiently localize and mitigate harmful instructions. Our extensive experiments, conducted across state-of-the-art models, confirm that AI reviewers are pervasively vulnerable. PaperGuard establishes the foundational benchmark, protocols, and actionable defense necessary to pioneer trustworthy, attack-resilient AI-assisted scholarly reviewing.

cs.CL↗
arXiv2026

EveLoad: Cognitive Workload Recognition from Event-Based Eye Movements

Cognitive workload monitoring is important for adaptive rehabilitation and assistive interfaces, where task difficulty, pacing, and feedback should be adjusted according to the user's cognitive state to avoid overload and under-challenge. Emerging extended reality and robot-assisted rehabilitation environments provide controllable training tasks, but they require unobtrusive sensing methods that can capture rapid ocular dynamics during interaction. Existing eye-movement-based cognitive workload recognition methods mainly rely on frame-based eye trackers, which often suffer from limited temporal resolution and degraded robustness under rapid eye movements. In contrast, event cameras provide microsecond-level temporal resolution, high dynamic range and low latency, making them suitable for capturing fine-grained ocular dynamics. Many previous studies rely on free-viewing or similar paradigms, where gaze locations can vary across tasks. As a result, models may learn associations between gaze-location distributions and cognitive workload, rather than workload-related eye movement characteristics themselves. In this work, we introduce EveLoad, which, to the best of our knowledge, is the first event-based eye-movement dataset with graded cognitive workload annotations, collected from 20 healthy participants under spatially constrained and task-driven conditions using a controlled N-back-guided fixation paradigm. Based on this dataset, we establish a benchmark for cognitive workload recognition with six workload levels and propose a learning framework that encodes spatiotemporal event representations. Experimental results show that our approach achieves an average subject-specific accuracy of 96.36% and 96.13% under mixed random split evaluation. These results suggest that event-based eye movements may provide a useful sensing pathway for future workload-aware rehabilitation.

cs.LG↗
arXiv2026

PieArena: Ranking and Profiling Language Agents in Realistic Negotiation Scenarios

We present an in-depth evaluation of LLMs' ability to negotiate, a central business task requiring strategic reasoning, theory of mind, and economic value creation. To do so, we introduce PieArena, a large-scale negotiation benchmark grounded in multi-agent interactions over realistic scenarios adapted from MBA negotiation courses at an elite business school. We evaluate language agents across three pairing regimes: mirror-play, cross-play, and human-LM play. We develop a ranking model for continuous negotiation payoffs that yields order-invariant, uncertainty-quantified leaderboards while correcting for systematic experimental asymmetries. We further study the effects of joint-intentionality agentic scaffolding and find asymmetric gains, with large improvements for mid- and lower-tier LMs and diminishing returns for frontier LMs. As calibration anchors, we collect human-human and human-LM negotiation data from trained business school students, finding that a representative frontier language agent (GPT-5) matches or exceeds this human baseline in our evaluation settings. Beyond deal outcomes, PieArena provides a multi-dimensional behavioral profile that reveals cross-model heterogeneity in instruction compliance, computation accuracy, as well as judge-assessed deception and reputation, illustrating the value of evaluation beyond outcome-only leaderboards.

cs.AI↗
arXiv2026

Repurposing Adversarial Perturbations for Continual Learning: From Defense to Active Alignment

In dynamic environments, large language models need to keep adapting to new tasks, but continual learning often suffers from forgetting, limited transfer, and vulnerability to adversarial perturbations. To address this, we present AdvCL, which repurposes adversarial perturbations as a geometric control signal for stable continual adaptation. AdvCL combines three plug-in modules: Intra-Smooth promotes local smoothness via small adversarial perturbations; Proto-Clip uses similarity clipping to prevent excessive alignment to current task prototype; and Inter-Align applies directional alignment toward previous task prototype to reduce representational gaps. Experiments show consistent gains in both standard performance and robustness, with lower forgetting and stronger transfer. We further analyze key mechanisms by quantifying the sensitivity of Intra-Smooth to perturbation settings and the effect of Inter-Align on task similarity and geometric distance. In summary, the modules provide complementary gains when combined, and each can also be integrated individually into diverse CL paradigms, including replay, regularization, and dynamic architectures, thereby offering a geometric control mechanism for continual learning.

cs.LG↗
arXiv2026

Projected Energy Correlators: Two-Loop Jet Functions and NNLL Resummation

We present the next-to-next-to-leading logarithmic (NNLL) collinear resummation of projected $N$-point energy correlators (ENCs) up to $N=6$, matched to fixed-order predictions at NLO, in both electron-positron annihilation and Higgs decay to gluons. The key new ingredient is the two-loop jet function for $N=4,5,6$, which we compute semi-analytically using Integration-by-Parts and differential equations. We further include the leading non-perturbative corrections for ENCs, described by two universal soft matrix elements $\overlineΩ_{1q},\overlineΩ_{1g}$ of order $Λ_{\rm QCD}$, whose evolution is governed by anomalous dimensions for $(N-1)$-point correlators. The matched distributions are compared with parton-shower simulations from Pythia8 and Herwig7, and we study the sensitivity of both the absolute spectra and their ratios to the two-point energy correlator under variations of $α_s$ and $\overlineΩ_{1q,1g}$. Our results show that higher-point projected energy correlators are now under quantitative control at NNLL accuracy, opening the door to future $α_s$ extractions with complementary systematics.

hep-ph↗
arXiv2026

DiagramBank: A Quality-Audited Dataset of Scientific Schematic Diagrams with Multi-Level Document Context

Scientific papers use schematic diagrams to communicate methods, workflows, and system structure, yet existing scientific-figure corpora often mix them with plots, screenshots, and photographs and rarely preserve document context. We introduce DiagramBank, a quality-audited dataset of 57,100 schematic diagrams curated from OpenReview-hosted AI/ML venues. Each record links a diagram image to its paper title, abstract, figure caption, in-text figure-reference spans, venue/year metadata, provenance fields, and filtering labels. DiagramBank is a reusable resource for scientific-document understanding, diagram retrieval, corpus analysis, and future benchmark construction. We describe its extraction and cascade-filtering pipeline, release schema, confidence-controlled views, dataset card, and indexing utilities. A manual blind audit of the released cascade-filtered records estimates 93.67% precision, and a separate CLIP threshold analysis characterizes the precision--coverage trade-off for simpler filtering views. We further provide lightweight metadata-indexing and authoring examples to illustrate downstream protocols without treating these utilities as standalone methods. The code is public at: https://github.com/csml-rpi/DiagramBank.

cs.IR↗
arXiv2026

Rethinking Side-Channel Analysis: Automated Discovery and Analysis of Side-Channel Leakage with LLM-Assisted Agents

Side-channel attacks exploit unintended information leakage from system behavior and continue to pose serious privacy risks in modern platforms. Despite extensive prior work, side-channel analysis remains largely manual and fragmented, typically assuming predefined target events and a fixed set of known channels. As systems and applications grow increasingly complex, several fundamental questions remain unanswered: which user or system events are sensitive in practice, how side channels associated with these events can be systematically discovered without exhaustive manual effort, and how their leakage can be analyzed at scale without prohibitive data collection and model training costs. To address these questions, we present SCAgent, an automated framework for side-channel risk analysis. To identify sensitive targets beyond manually specified events, SCAgent performs agent-driven system exploration guided by LLM-based semantic reasoning. To systematically discover side channels while mitigating the risk of LLM hallucination, it reasons over system documentation and incorporates explicit verification to enforce semantic consistency, threat-model feasibility, and per-channel usability. To enable scalable analysis under limited data, SCAgent adopts a few-shot learning paradigm based on foundation models, avoiding the need to train bespoke models for each channel--event pair. To bridge the gap between raw time-series side-channel signals and tabular foundation models, SCAgent further introduces a time-shift--robust feature extraction layer that enables effective downstream analysis. We instantiate SCAgent on iOS as a first step, focusing on OS-level side channels observable by unprivileged applications. Our evaluation spans standard benchmarks such as foreground app and website fingerprinting, as well as newly identified sensitive in-app activities in popular applications.

cs.CR↗
arXiv2026

Not All Prefills Are Equal: PPD Disaggregation for Multi-turn LLM Serving

Prefill-Decode (PD) disaggregation has become the standard architecture for modern LLM inference engines, which alleviates the interference of two distinctive workloads. With the growing demand for multi-turn interactions in chatbots and agentic systems, we re-examined PD in this case and found two fundamental inefficiencies: (1) every turn requires prefilling the new prompt and response from the last turn, and (2) repeated KV transfers between prefill and decode nodes saturate the bandwidth, leading to high latency and even service degradation. Our key insight is that not all prefill operations are equally disruptive: append-prefill, which processes only the new input tokens while reusing cached KV states, incurs an order-of-magnitude smaller decoding slowdown than full prefill. This motivates routing append-prefill to decode nodes locally. However, through comprehensive analysis, we show that no single fixed routing strategy satisfies all Service Level Objectives (SLOs) simultaneously. Based on this insight, we propose Prefill Prefill-capable Decode (PPD) disaggregation, a dynamic routing system that decides when to process Turn 2+ requests locally on decode nodes using cached KV states. PPD adapts to varying SLOs via configurable weights and seamlessly integrates with traditional PD deployments. With extensive evaluations, we show that PPD reduces Turn 2+ time-to-first-token (TTFT) by $\sim$68\% while maintaining competitive time-per-output-token (TPOT), effectively alleviating KV transfer congestion under high load. PPD provides a flexible and efficient paradigm for multi-turn LLM serving.

cs.NI↗
arXiv2026

Characterizing Trust Boundary Vulnerabilities in TEE Containers: An Empirical Study

Trusted Execution Environments (TEEs) have become a cornerstone of confidential computing, attracting significant attention from academia and industry. To support secure and scalable application deployment on confidential clouds, TEE containers (Tcons) have been introduced as middleware to shield applications from malicious operating systems and orchestration layers while preserving usability. In this paper, we present the first comprehensive analysis of Tcons, focusing on three critical layers: OS interfaces, encrypted I/O, and orchestration mechanisms. To enable systematic evaluation, we design TBouncer, an automated analyzer that precisely exercises and benchmarks Tcon isolation boundaries. Our study uncovers fundamental flaws in existing Tcons, leading to exploitable vulnerabilities such as code execution, denial-of-service, and information leakage. In total, we identify six attack vectors, twelve new bugs, and three CVEs. These findings provide new insights into the underestimated attack surface of Tcons and highlight key directions for building more secure and trustworthy container solutions.

cs.CR↗
arXiv2026

M$^3$-ACE: Rectifying Visual Perception in Multimodal Math Reasoning via Multi-Agentic Context Engineering

Multimodal large language models have recently shown promising progress in visual mathematical reasoning. However, their performance is often limited by a critical yet underexplored bottleneck: inaccurate visual perception. Through systematic analysis, we find that the most failures originate from incorrect or incomplete visual evidence extraction rather than deficiencies in reasoning capability. Moreover, models tend to remain overly confident in their initial perceptions, making standard strategies such as prompt engineering, multi-round self-reflection, or posterior guidance insufficient to reliably correct errors. To address this limitation, we propose M3-ACE, a multi-agentic context engineering framework designed to rectify visual perception in multimodal math reasoning. Instead of directly aggregating final answers, our approach decouples perception and reasoning by dynamically maintaining a shared context centered on visual evidence lists. Multiple agents collaboratively contribute complementary observations, enabling the system to expose inconsistencies and recover missing perceptual information. To support stable multi-turn collaboration, we further introduce two lightweight tools: a Summary Tool that organizes evidence from different agents into consistent, complementary, and conflicting components, and a Refine Tool that filters unreliable samples and guides iterative correction. Extensive experiments demonstrate that M3-ACE substantially improves visual mathematical reasoning performance across multiple benchmarks. Our method establishes new state-of-the-art results 89.1 on the MathVision benchmark and achieves consistent improvements on other related datasets, including MathVista and MathVerse. These results highlight the importance of perception-centric multi-agent collaboration for advancing multimodal reasoning systems.

cs.AI↗
arXiv2026

When Does Divide and Conquer Work for Long Context LLM? A Noise Decomposition Framework

We investigate the challenge of applying Large Language Models (LLMs) to long texts. We propose a theoretical framework that distinguishes the failure modes of long context tasks into three categories: cross-chunk dependence (task noise), confusion that grows with context size (model noise), and the imperfect integration of partial results (aggregator noise). Under this view, we analyze when it is effective to use multi-agent chunking, i.e., dividing a lengthy sequence into smaller chunks and aggregating the processed results of each chunk. Our experiments on tasks such as retrieval, question answering, and summarization confirm both the theoretical analysis and the conditions that favor multi-agent chunking. By exploring the accelerated decay of model fidelity with input length, we also explain why, for large inputs, a weaker model configured with chunk-based processing can surpass a more advanced model like GPT4o applied in a single shot. Overall, we present a principled understanding framework and our results highlight a direct pathway to handling long contexts in LLMs with carefully managed chunking and aggregator strategies.

cs.CL↗
arXiv2026

Joint Optimization of Flexible Antenna Array Shape and Beamforming for Secure Communication

Flexible antenna arrays (FAAs) can physically reshape their geometry to add new spatial degrees of freedom, whereas transmit beamforming adjusts the complex element weights to electronically steer and shape the array's radiation pattern, thereby significantly improving communication performance. This paper is the first to explore the integration of FAA geometry control and beamforming for physical layer security enhancement, where a base station equipped with an FAA communicates with a legitimate user in the presence of passive eavesdroppers. To safeguard confidential transmissions, we formulate a new secrecy rate maximization problem that jointly optimizes the transmit beamforming vector and a continuous FAA shape control parameter. Due to the non convex nature of the problem, an alternating optimization algorithm is developed to decompose the joint design into tractable subproblems, which are solved iteratively to refine both the FAA geometry and beamforming strategy. Simulation results confirm that the proposed joint optimization framework significantly outperforms conventional fixed shape or beamforming only schemes, demonstrating the potential of FAA enabled reconfigurability for secure wireless communications.

eess.SP↗
arXiv2026

Ternary Spiking Neural Networks Enhanced by Complemented Neurons and Membrane Potential Aggregation

Spiking Neural Networks (SNNs) are promising energy-efficient models and powerful framworks of modeling neuron dynamics. However, existing binary spiking neurons exhibit limited biological plausibilities and low information capacity. Recently developed ternary spiking neuron possesses higher consistency with biological principles (i.e. excitation-inhibition balance mechanism). Despite of this, the ternary spiking neuron suffers from defects including iterative information loss, temporal gradient vanishing and irregular distributions of membrane potentials. To address these issues, we propose Complemented Ternary Spiking Neuron (CTSN), a novel ternary spiking neuron model that incorporates an learnable complemental term to store information from historical inputs. CTSN effectively improves the deficiencies of ternary spiking neuron, while the embedded learnable factors enable CTSN to adaptively adjust neuron dynamics, providing strong neural heterogeneity. Furthermore, based on the temporal evolution features of ternary spiking neurons' membrane potential distributions, we propose the Temporal Membrane Potential Regularization (TMPR) training method. TMPR introduces time-varying regularization strategy utilizing membrane potentials, furhter enhancing the training process by creating extra backpropagation paths. We validate our methods through extensive experiments on various datasets, demonstrating remarkable performance advances.

cs.NE↗
arXiv2026

Enhancing LLM-Based Data Annotation with Error Decomposition

Large language models offer a scalable alternative to human coding for data annotation tasks, enabling the scale-up of research across data-intensive domains. While LLMs are already achieving near-human accuracy on objective annotation tasks, their performance on subjective annotation tasks, such as those involving psychological constructs, is less consistent and more prone to errors. Standard evaluation practices typically collapse all annotation errors into a single alignment metric, but this simplified approach may obscure different kinds of errors that affect final analytical conclusions in different ways. Here, we propose a diagnostic evaluation paradigm that incorporates a human-in-the-loop step to separate task-inherent ambiguity from model-driven inaccuracies and assess annotation quality in terms of their potential downstream impacts. We refine this paradigm on ordinal annotation tasks, which are common in subjective annotation. The refined paradigm includes: (1) a diagnostic taxonomy that categorizes LLM annotation errors along two dimensions: source (model-specific vs. task-inherent) and type (boundary ambiguity vs. conceptual misidentification); (2) a lightweight human annotation test to estimate task-inherent ambiguity from LLM annotations; and (3) a computational method to decompose observed LLM annotation errors following our taxonomy. We validate this paradigm on four educational annotation tasks, demonstrating both its conceptual validity and practical utility. Theoretically, our work provides empirical evidence for why excessively high alignment is unrealistic in specific annotation tasks and why single alignment metrics inadequately reflect the quality of LLM annotations. In practice, our paradigm can be a low-cost diagnostic tool that assesses the suitability of a given task for LLM annotation and provides actionable insights for further technical optimization.

cs.CL↗
arXiv2026

Evaluating 21st-Century Competencies in Postsecondary Curricula with Large Language Models: Performance Benchmarking and Reasoning-Based Prompting Strategies

The growing emphasis on 21st-century competencies in postsecondary education, intensified by the transformative impact of generative AI, underscores the need to evaluate how these competencies are embedded in curricula and how effectively academic programs align with evolving workforce and societal demands. Curricular Analytics, particularly recent generative AI-powered approaches, offer a promising data-driven pathway. However, analyzing 21st-century competencies requires pedagogical reasoning beyond surface-level information retrieval, and the capabilities of large language models in this context remain underexplored. In this study, we extend prior curricular analytics research by examining a broader range of curriculum documents, competency frameworks, and models. Using 7,600 manually annotated curriculum-competency alignment scores, we assess the informativeness of different curriculum sources, benchmark general-purpose LLMs for curriculum-to-competency mapping, and analyze error patterns. We further introduce a reasoning-based prompting strategy, Curricular CoT, to strengthen LLMs' pedagogical reasoning. Our results show that detailed instructional activity descriptions are the most informative type of curriculum document for competency analytics. Open-weight LLMs achieve accuracy comparable to proprietary models on coarse-grained tasks, demonstrating their scalability and cost-effectiveness for institutional use. However, no model reaches human-level precision in fine-grained pedagogical reasoning. Our proposed Curricular CoT yields modest improvements by reducing bias in instructional keyword inference and improving the detection of nuanced pedagogical evidence in long text. Together, these findings highlight the untapped potential of institutional curriculum documents and provide an empirical foundation for advancing AI-driven curricular analytics.

cs.CY↗
arXiv2026

Collaborative Multi-Agent Test-Time Reinforcement Learning for Reasoning

Multi-agent systems have evolved into practical LLM-driven collaborators for many applications, gaining robustness from diversity and cross-checking. However, multi-agent RL (MARL) training is resource-intensive and unstable: co-adapting teammates induce non-stationarity, and rewards are often sparse and high-variance. Therefore, we introduce \textbf{Multi-Agent Test-Time Reinforcement Learning (MATTRL)}, a framework that injects structured textual experience into multi-agent deliberation at inference time. MATTRL forms a multi-expert team of specialists for multi-turn discussions, retrieves and integrates test-time experiences, and reaches consensus for final decision-making. We also study credit assignment for constructing a turn-level experience pool, then reinjecting it into the dialogue. Across challenging benchmarks in medicine, math, and education, MATTRL improves accuracy by an average of 3.67\% over a multi-agent baseline, and by 8.67\% over comparable single-agent baselines. Ablation studies examine different credit-assignment schemes and provide a detailed comparison of how they affect training outcomes. MATTRL offers a stable, effective and efficient path to distribution-shift-robust multi-agent reasoning without tuning.

cs.AI↗
arXiv2026

Temporal Regularization Training: Unleashing the Potential of Spiking Neural Networks

Spiking Neural Networks (SNNs) have received widespread attention due to their event-driven and low-power characteristics, making them particularly effective for processing neuromorphic data. Recent studies have shown that directly trained SNNs suffer from severe temporal gradient vanishing and overfitting issues, which fundamentally constrain their performance and generalizability. This paper unveils a temporal regularization training (TRT) memthod, designed to unleash the generalization and performance potential of SNNs through a time-decaying regularization mechanism that prioritizes early timesteps with stronger constraints. We perform theoretical analysis to reveal TRT's ability on mitigating the temporal gradient vanishment. To validate the effectiveness of TRT, we conduct experiments on both static image datasets and dynamic neuromorphic datasets, perform analysis of their results, demonstrating that TRT can effectively mitigate overfitting and help SNNs converge into flatter local minima with better generalizability. Furthermore, we establish a theoretical interpretation of TRT's temporal regularization mechanism by analyzing the temporal information dynamics inside SNNs. We track the Fisher information of SNNs during training process, showing that Fisher information progressively concentrates in early timesteps. The time-decaying regularization mechanism implemented in TRT effectively guides the network to learn robust features in early timesteps with rich information, thereby leading to significant improvements in model generalization.

cs.NE↗
arXiv2025

Invariant fractocohesive length in thermally aged elastomers

The fractocohesive length - the ratio between fracture toughness and work-to-fracture - provides a material-specific length scale that characterizes the size-dependent fracture behavior of pristine elastomers. However, its relevance to thermally aged materials, where both toughness and work of fracture degrade dramatically, remains unexplored. Here, we demonstrate that despite severe thermal embrittlement, the fractocohesive length remains invariant throughout thermal aging, independent of temperature or duration. We verify this invariance experimentally for two elastomer systems (Styrene Butadiene Rubber and Silicone Rubber) at multiple aging temperatures for aging times up to eight weeks. This finding bridges a critical gap in fracture mechanics of aged polymers: while the evolution of work-to-fracture can be predicted from well-established constitutive models that track network changes (crosslink density and chain scission), the evolution of fracture toughness has lacked predictive frameworks. The invariance of fractocohesive length enables direct calculation of fracture toughness at any aging state from the predicted work of fracture, eliminating the need for extensive fracture testing on aged elastomers and providing a crucial missing link for computational fracture predictions in aged elastomeric components.

cond-mat.soft↗