arXiv subjects

Jun Wu

Publications and source records attributed to Jun Wu.

At least 19 recordsLinked to original sources

arXiv2026

Fluid Antennas Enabled UAV Communication Systems: Fundamentals, Challenges, and Future Directions

Low-altitude wireless networks (LAWNs) require uncrewed aerial vehicle (UAV) communication systems to adapt to large-scale propagation changes and fast local channel fluctuations. Fluid antennas (FAs) provide an additional degree of freedom (DoF) by reconfiguring antenna positions within a compact aperture. This article investigates FA-enabled UAV communication systems from the perspectives of fundamentals, design challenges, and future research directions. We first introduce a dual-mobility perspective, in which UAV trajectory planning provides macro-mobility for large-scale air-to-ground geometry adaptation, and FA position reconfiguration provides micro-mobility for fine-grained local channel adaptation. We then identify key design challenges arising from the high computational complexity and information uncertainty in the joint optimization of UAV trajectory, FA positions, and beamforming under mobility, energy, hardware, and security constraints. To address these challenges, we present a hybrid offline-online optimization framework, where prior information such as environment maps, user mobility patterns, and statistical channel state information (CSI) supports long-term planning, while real-time information such as instantaneous CSI and UAV state enables fast online correction. A case study on FA-enabled UAV-assisted vehicular communications demonstrates the effectiveness of the proposed framework. Finally, we discuss future directions of FA-enabled UAV communication in integrated sensing and communication (ISAC), secure communication, mobile edge computing (MEC), and aerial relaying.

eess.SP↗

arXiv2026

Strategic Calibration in LLM Multi-Agent Systems: Audited Pivot Scoring for Truthful Elicitation

In large language model (LLM)-based multi-agent systems, reported confidence increasingly determines the allocation of scarce downstream attention, computation, and decision influence. This makes reporting strategic: an agent may overstate the relevance of its contribution even when its underlying belief is well calibrated. Existing calibration methods largely assume passive reporting and therefore do not address strategic incentives created when confidence determines downstream allocation. We formulate strategic calibration as a reporting-layer incentive problem and propose the Audited Pivot Scoring Mechanism (APSM). APSM offsets platform-controlled allocation benefits, uses inverse-propensity-weighted sparse audits, charges the Clarke pivot externality, and applies a strictly proper score. Under the stated model assumptions, truthful reporting is the unique expected-utility best response. We also characterize approximate execution and the audit-variance trade-off. Calculator-free experiments show that LLM agents inflate confidence under Top-$K$ incentives. With a shared agent-local utility interface, LLM agents reduce distortion under both Compensated Proper Scoring (CPS) and APSM. CPS achieves higher exact allocation match, whereas APSM yields lower Brier loss and numerically fewer cutoff crossings. Experiments further show that APSM enlarges the utility margin against allocation-changing deviations by $1.93\times$ and improves robustness under approximate execution. These results support incentive-aware elicitation as a distinct layer for reliable multi-agent coordination.

cs.GT↗

arXiv2026

Affine and cyclotomic Brauer categorification via $\imath$-Kac--Moody $2$-categories: the integral type $\operatorname{AIII}$ case

Continuing our work on the half-integral case, we construct generalized nilpotent $2$-representations of the integral quasi-split type $\operatorname{AIII}$ $\imath$-Kac--Moody $2$-category from affine Brauer actions with dot spectrum exactly $\mathbb Z$ on locally Schurian $\mathbb C$-linear categories. Conversely, nilpotent $2$-representations on such categories yield affine Brauer actions with spectrum contained in $\mathbb Z$. At the fixed vertex $0$, we directly define corrected generators and verify the nil-Brauer and mixed relations. For $\mathbf u$-admissible integral parameters at arbitrary positive level, we prove an isomorphism between the locally unital algebras of the cyclotomic Brauer category and a cyclotomic quotient of the principal $2$-representation. This gives compatible $\mathbb Z$-gradings on the category and its endomorphism algebras, providing a Brauer analogue of the graded realization of type $A$ cyclotomic Hecke algebras. Over algebraically closed fields of odd characteristic $p$, we also obtain both constructions with spectrum contained in $\mathbb F_p$, for the mixed affine datum restricted to its spectral weight subgroup.

math.RT↗

arXiv2026

TaQFormer: A Treatment-as-Query Transformer with Data-Fusion for Multidimensional Treatment Modeling in Coupon Allocation

We study coupon allocation on an online marketplace offering multiple service categories. Issuing coupons requires choosing a vector of category-specific discounts, accounting for both their direct effects and cross-category demand shifts. We present TaQFormer, a treatment-as-query Transformer for this multidimensional treatment setting. Each treatment coordinate is represented as a query token, allowing cross-attention to estimate the response to the full treatment vector. To combine large but policy-biased observational logs with a smaller stream of randomized controlled trial data, we use a two-stage pipeline: observational pretraining learns general outcome patterns, followed by LoRA fine-tuning on RCT data with a request-specific fusion gate. We evaluate the resulting system through the checks required for production, including offline uplift metrics, treatment-parallel serving, latency monitoring, and a 14-day online A/B test. At essentially the same subsidy cost rate, \model increases business value per user by 1.26\% over the production baseline and significantly outperforms both the production baseline and the RCT-only model, while meeting the latency budget. It was subsequently rolled out to all marketplace traffic and has operated at full scale for several months as of October 2026. The deployment illustrates how multidimensional treatment modeling and observational pretraining, combined with adaptation using RCT data, can support effective joint discount decisions in production.

cs.LG↗

arXiv2026

Diagnostic Analysis for Large Language Model Reasoning

Reasoning structures (e.g., chains and graphs) are increasingly used for probabilistic inference, factuality verification, and uncertainty estimation, but their structural reliability is rarely evaluated directly. We test whether this reliability holds through a proposed two-stage diagnostic evaluation in open-ended multi-hop QA. Stage 1 tests different capabilities required to construct reasoning graphs, showing that graph reliability varies with topology, prompt design, and model family. Stage 2 uses uncertainty estimation as a downstream test case. Generated-edge experiments show that graph-based computation is sensitive to upstream graph errors, with performance dropping relative to the gold-graph setting. Gold-graph experiments further show that graph structure provides little benefit when used only as prompt context, but improves the separation of correct from incorrect answers by confidence score and enables node-level diagnostics.

cs.CL↗

arXiv2026

When Messages Go Wrong: Understanding Communication Faults in LLM-Based Multi-Agent Systems

LLM-based multi-agent systems (LLM-MAS) use specialized agents that exchange messages to accomplish complex tasks. Their reliance on networked protocols, remote tools, and model endpoints exposes communication to message loss, delay, and context expiration. Because agents often exchange open-ended natural language, anomalies can alter task-relevant information while leaving messages processable, allowing semantic errors to propagate through subsequent interactions. Systematic studies of faults in inter-agent communication remain limited. This paper presents a large-scale empirical study of how communication faults manifest, propagate, and affect task outcomes in LLM-MAS. We construct a cross-layer taxonomy linking lower-stack disturbances to application-visible faults and develop a fault-injection and tracing pipeline. Our evaluation covers more than 29,000 executions across 100 tasks, five domains, three workflow topologies, and three model configurations. We find that 20.7% of failures with complete detection records occur without an immediate failure signal, reaching 76.2% for partial-content faults. Flat maintains 53.0% task success under communication faults, compared with 3.3% and 3.2% for Sequential and Hierarchical, respectively. This advantage primarily arises when an alternative information path preserves the evidence required by downstream agents. We also evaluate execution-side and evidence protection mechanisms and find that they recover different failure cases. Evidence recovery depends on whether trustworthy information remains accessible through an unaffected path. These observations provide empirical insights into communication-fault propagation, workflow resilience, and protection design for LLM-MAS.

cs.SE↗

arXiv2026

PointWalker: Mapless Quadruped Locomotion Directly From Raw LiDAR Point Clouds

Quadruped robots deployed for inspection and field missions must cross heterogeneous terrain, from stairs, slopes, and gaps to high platforms, stepping stones, and balance beams, where safe support is sparse or see-through. Existing methods interpose a terrain map or a depth image between sensing and control. Maps are costly to maintain, corrupted by odometry drift and outliers, and misled by see-through gratings, while depth cameras degrade with lighting and surface material. In this paper, we present PointWalker, a mapless framework whose single reinforcement learning policy directly consumes raw point clouds from four light detection and ranging (LiDAR) sensors. Voxelized points are encoded by linear self-attention and fused with proprioception through cross-attention, so terrain geometry reaches the policy without maps or foothold estimation. One proximal policy optimization (PPO) run, with terrain-aware rewards and dual-channel simulation of see-through gratings, yields one policy for all seven terrain classes with a 90.1% mean success rate in simulation. Deployed zero-shot on a Deep Robotics X30 at 50 Hz onboard, it crosses 0.8-m gaps, climbs 0.8-m platforms and 45° steel-grating stairs, carries a 50-kg payload, and completes a 2.5-km mountain route.

cs.RO↗

arXiv2026

Detecting Unseen Jailbreak Sources: A Multi-Source Conformal Detection Perspective

Jailbreak defenses for large language models are usually calibrated on a fixed set of known attack methods. However, new attack methods keep appearing after deployment, and existing detectors are observed to degrade under the new attacks (Piet et al., 2025). This paper studies how to detect, from a stream of attack prompts, that the prompts come from an attack source outside the known ones. We formulate it as a sequential test whose null hypothesis is that the stream comes from one of the known sources. Methodology-wise, we compute a conformal p-value against each known source, and reject the null only when the evidence against every known source is large. Using these source-wise p-values, we build Persist-3 and analyze its stationary false-alarm probability and detection power. The technical contribution of Persist-3 lies in a detection guarantee that uses only a finite window of recent observations. Our theoretical analysis reveals that its detection probability approaches one exponentially fast as the window grows, and thus a short window suffices to detect unseen sources that are well separated from all known sources. Empirically, we validate Persist-3 on top of two recent jailbreak defenses, MTK (Zhang et al., 2026) and Prompt Guard 2 (Chennabasappa et al., 2025), and observe a significant defense enhancement.

stat.AP↗

arXiv2026

Asymptotic Entanglement Hiding under Stabilizer Restrictions

Entanglement is central to quantum information processing, while stabilizer operations underpin fault-tolerant quantum computation. We ask how much entanglement remains visible or distillable under stabilizer restrictions. We quantify stabilizer-visible entanglement by restricting the measured relative entropy of entanglement to stabilizer measurements, thereby obtaining converse bounds on entanglement distillation under stabilizer operations. We demonstrate magic-free asymptotic entanglement hiding: we construct explicit convex mixtures of pure stabilizer states on $N$ qutrits per party whose unrestricted visible entanglement and LOCC-distillable entanglement both grow as $Ω(N/\log N)$, while their stabilizer-visible and stabilizer-distillable entanglement vanish as $N\to\infty$. Thus, an unbounded amount of LOCC-distillable entanglement carried by stabilizer states can become asymptotically invisible and undistillable under stabilizer restrictions. We further prove that stabilizer-visible entanglement is $O(1)$ with high probability for Haar-random pure states despite extensive unrestricted visibility, and vanishes uniformly over entangled Werner states as the local dimension grows through odd primes. These results reveal a fundamental separation between entanglement and magic as resources, exposing intrinsic limits on entanglement extraction using stabilizer operations.

quant-ph↗

arXiv2026

GUIDE: Goal-Initialized Directional Understanding for End-to-End Legged Navigation

End-to-end reinforcement learning (RL) has shown strong potential for legged robot navigation, yet existing approaches commonly rely on continuously updated robot-to-goal states from external state estimation modules, leaving part of the navigation problem outside the learned policy. In this work, we seek to push the limits of end-to-end sim-to-real RL navigation by investigating whether a legged robot can internally maintain the spatial context required for long-horizon navigation. To this end, we study goal-initialized navigation, where the goal is provided only once at the beginning of an episode, with no subsequent external relative-goal updates. We present GUIDE, an end-to-end RL framework that jointly learns navigation and internal directional awareness from onboard observations. GUIDE leverages multi-frequency proprioceptive history to capture egomotion and auxiliary spatial-anchor prediction to maintain task-relevant spatial states, while temporal depth observations provide local environmental geometry. The navigation policy is trained entirely in simulation and transferred zero-shot to the real world. Across cluttered environments and structured mazes, GUIDE reliably avoids obstacles, escapes dead ends, and reaches distant goals using only onboard sensing. These results demonstrate that robust sim-to-real legged navigation can be achieved without continuously providing external robot-to-goal estimation, opening a promising direction for future research on more self-contained end-to-end navigation.

cs.RO↗

arXiv2026

UniPoint: Unified Point-Level Sensor Fusion for Humanoid Locomotion Across Challenging Terrains

Open-world deployment requires humanoid robots to cross highly heterogeneous terrain safely, with perception that simultaneously provides wide coverage, local accuracy, and redundancy against sensor failure. Existing approaches struggle to satisfy all three: one forward depth camera or nearby height sampling covers too little; odometry-corrected elevation maps drift under aggressive motion and miss thin vertical structures; image-level encoding costs grow with camera count. We present UniPoint, a humanoid whole-body locomotion framework built on multi-source point-level sensor fusion. Measurements from a 360° light detection and ranging (LiDAR) sensor and two depth cameras are early-fused into one base-frame point set. Voxelization resamples it to a fixed number of tokens encoded by linear self-attention and proprioception-queried cross-attention, decoupling forward cost from sensor count. The point set retains standing thin barriers; a single-modality failure removes only part of the tokens, so the policy degrades gracefully. A single training run with terrain-aware rewards, perception-degradation injection, and domain randomization produces one policy for all eight terrain types, deployed on an onboard RK3588 without fine-tuning. On a DR02 humanoid, 20 trials at each of nine real-world settings over seven terrain types validate the policy on 70-cm-high platforms, 100-cm gaps, thin barriers, and sparse or narrow footholds; it also generalizes zero-shot outdoors.

cs.RO↗

arXiv2026

SoK: Intent-Oriented Systematization of Multi-Turn LLM Jailbreaks

Large Language Models (LLMs) are increasingly deployed in interactive settings, where user intent commonly unfolds through multi-turn dialogue. Multi-turn jailbreaks exploit this pattern by advancing a harmful intent across turns, so that no single message exposes the full objective. However, existing work treats these attacks as a loose collection of prompt patterns and does not analyze how the adversary organizes and advances harmful intent across an interaction. We develop a four-part, intent-oriented taxonomy that organizes multi-turn jailbreaks by adversarial intent structure. Through controlled ablations, we find that effectiveness is driven by how deliberately intent is organized across turns rather than by context length or query count. We further show that the way intent is organized determines the level at which it becomes detectable, pushing the required detection surface outward from the turn level to the session level to the cross-session level. These findings indicate that turn-local safety mechanisms are structurally insufficient and that single-point evaluation overlooks how intent is organized, motivating evaluation protocols aligned to the level at which harmful intent becomes observable. The code is available at: https://github.com/SiyuanLi00/INTACT.

cs.CR↗

arXiv2026

Connecting heterogeneous dynamics with local entropy

Establishing a robust and physically interpretable link between static structure and heterogeneous relaxation dynamics remains a fundamental challenge in glass physics. Here, we introduce a weighted pair-entropy descriptor based on the conventional two-body excess entropy. For this, we multiply the integrand used to calculate the excess entropy by a weight function that is directly related to the length scale of the pair correlation function. This multiplication does not affect the contribution of the short range order to the local excess entropy, but allows to take into account the structure present on intermediate distances, i.e., the medium-range order. For a canonical two-dimensional Lennard-Jones glass former, the resulting descriptor exhibits a strong correlation with particle-level dynamical propensity at long times (multiples of the alpha-relaxation time), with a maximum structure--dynamics correlation reaching about 0.9, substantially outperforming the predictive power of the conventional local pair excess entropy. These results demonstrate that incorporating a physically motivated structural length scale into entropy-based descriptors markedly enhances their predictive power while preserving physical interpretability. Our findings provide a simple and general framework for investigating structure--dynamics correlations in glass-forming systems.

cond-mat.dis-nn↗

arXiv2026

Toward Intelligent Skies: Signal Processing and AI Foundations of Low-Altitude Wireless Networks

The rapid growth of low-altitude aerial services and applications, driven by uncrewed aerial vehicles (UAVs), calls for a new class of digital infrastructure beyond conventional terrestrial networks. The low-altitude wireless network (LAWN) has been proposed as dynamically reconfigurable three-dimensional architectures that integrate aerial and ground nodes to provide connectivity, sensing, and control in open, safety-critical airspace. This tutorial presents a comprehensive treatment of LAWNs from the joint perspectives of artificial intelligence (AI) and signal processing. We first review the historical evolution and architectural foundations of LAWNs, introducing altitude-based layers and functional planes, and summarizing the regulatory and standardization landscape. Building on this system view, we then discuss signal processing fundamentals for LAWNs, including 3D channel and system models, performance metrics, waveform and receiver design, localization and tracking, and multi-functionality co-design. Next, we survey AI techniques for LAWNs, covering discriminative and generative models for perception, control, resource management, and security, as well as emerging paradigms such as foundation models, large language models, and digital twins for mission planning and closed-loop optimization. To illustrate AI-signal processing integration in practice, we provide a case study of an AI-driven multi-tier LAWN with hybrid satellite, high-altitude, and ground nodes. The tutorial concludes by outlining key research challenges in architecture design, signal processing-AI co-design, safety and security, experimentation, and standardization, and by highlighting opportunities for LAWNs to evolve into dependable, AI-native infrastructure for the intelligent skies.

eess.SP↗

arXiv2026

Stateful Cooperative Agents Safeguarding LLMs Against Evolving Multi-Turn Attacks

As LLMs become increasingly integrated into complex applications, their vulnerability to adversarial attacks has raised significant concerns. However, existing defenses remain reactive in nature. This limitation makes it difficult for them to counter sophisticated threats, as adversaries continuously adjust their strategies across multi-turn interactions. In this paper, we present a proactive defense framework for securing LLMs against evolving multi-turn adversarial attacks that combines disruption, misdirection, and adaptation across successive interaction turns. In particular, it employs a cooperative multi-agent architecture in which specialized agents execute complementary defense strategies. These strategies include controlled response pacing to increase attack costs, strategically ambiguous outputs to mislead adversaries into ineffective strategies, and forensic analysis of interaction logs to identify attack patterns and refine defenses. These agents are coordinated by an adaptive mechanism that dynamically adjusts the defense strategy in response to escalating threats. To facilitate comprehensive evaluation, we present the EMRA dataset designed to simulate evolving strategies across multi-turn attacks, including 5,200 adversarial samples across eight attack types. Experimental results on EMRA across multiple LLM backbones show that the proposed framework reduces ASR by 69% on average relative to evaluated state-of-the-art baselines. Beyond suppressing harmful outputs, it sustains deceptive engagement, achieving an average DR more than six times that of the strongest baselines and increasing attacker-token consumption by 198.83% on average relative to evaluated baselines. Code and dataset are available at https://github.com/SiyuanLi00/CoopGuard.

cs.CR↗

arXiv2026

CReF: Cross-modal and Recurrent Fusion for Depth-conditioned Humanoid Locomotion

Stable traversal over geometrically complex terrain increasingly requires exteroceptive perception, yet prior perceptive humanoid locomotion methods often remain tied to explicit geometric abstractions, either by mediating control through robot-centric 2.5D terrain representations or by shaping depth learning with auxiliary geometry-related targets. While effective, these approaches introduce additional map-construction procedures or multi-stage skill-transfer processes beyond direct depth-to-control learning. We propose CReF (Cross-modal and Recurrent Fusion), a single-stage depth-conditioned humanoid locomotion framework that learns locomotion-relevant features directly from raw forward-facing depth without explicit geometric intermediates. CReF couples proprioception and depth tokens through proprioception-queried cross-modal attention, fuses the resulting representation with a gated residual fusion block, and performs temporal integration with a Gated Recurrent Unit (GRU) regulated by a highway-style output gate for state-dependent blending of recurrent and feedforward features. To further improve terrain interaction, we introduce a terrain-aware foothold placement reward that extracts supportable foothold candidates from foot-end point-cloud samples and rewards touchdown locations that lie close to the nearest supportable candidate. Experiments in simulation and on a physical humanoid demonstrate robust traversal over diverse terrains and effective zero-shot transfer to real-world scenes containing handrails, hollow pallet assemblies, severe reflective interference, and visually cluttered outdoor surroundings.

cs.RO↗

arXiv2026

Evidence-in-the-Loop: Trace-Driven Optimization for Customer-Service LLM Agents

Production customer-service bots must improve answer quality across iterative releases, yet large language models must not bypass evidence boundaries, policy rules, or human-handoff safeguards. We present an \textbf{Evidence-Grounded Customer-Service Agent Workflow} deployed in a real-world customer-service setting. BM25 recall, issue-title-vector recall, issue-description-vector recall, weighted RRF fusion, and cross-encoder reranking construct grounded FAQ evidence for controlled LLM decisions. Policy-guided orchestration then combines this RAG evidence with scenario-specific rule evidence, conversation memory, and clarification state inside a fixed LangGraph DAG~\cite{langgraph2024}. The paper contributes three reusable deployment patterns: \textbf{hybrid RAG evidence construction}, where multi-channel retrieval and reranking produce auditable FAQ candidates; \textbf{evidence-grounded issue/action decision}, where an Evidence-Grounded Decision Module selects an issue/action from typed FAQ evidence and scenario-specific rule evidence; and \textbf{trace-driven RAG and reranker improvement}, where traces diagnose whether failures come from recall, ranking, final candidate selection, clarification, rule-derived evidence, or action policy, and where reranker fine-tuning is evaluated not only for in-domain gain but also for forgetting risk.

cs.IR↗

arXiv2026

Identifiability of Partial-Mastery Cognitive Diagnostic Models

Partial-mastery (PM) cognitive diagnostic models (CDMs) extend traditional CDMs by replacing binary latent attribute mastery indicators with continuous mastery scores for multiple latent attributes. In PM-CDMs, each subject is characterized by a fixed continuous latent mastery vector, from which item-specific binary attribute profiles are independently generated. This formulation provides a bridge between classical CDMs and continuous latent variable models. Despite growing interest in PM-CDMs, their identifiability properties remain unexplored. In this work, we establish the first identifiability results for PM-CDMs. We derive sufficient conditions for identifiability that are direct analogues of established conditions for traditional CDMs. To develop the main argument, we use symbolic computation on a minimal example with five items and two latent attributes to show that the Jacobian of the model parameterization is generically nonzero. Combining tools from real analysis and algebraic statistics, we prove that this local property implies generic finite-to-one identifiability of the item parameters and the marginal distributions of the relevant latent attributes. We further show that if the $Q$-matrix contains such identifiable local structures for all attribute pairs, identifiability extends to the full PM-CDM. These findings provide a rigorous theoretical foundation for estimation and inference in partial-mastery cognitive diagnostic models.

math.ST↗