Search arXiv⌕ Search

arXiv subjects

Han Zhang

Publications and source records attributed to Han Zhang.

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

Acmite: Mitigating Gender Bias in LLMs through Concept-Guided Mutual Information

Large language models (LLMs) can reproduce social stereotypes from their training data, motivating extensive research on model debiasing. However, existing methods often rely on explicit biased examples or predefined group-term substitutions, making them sensitive to wording and less effective at capturing stereotype concepts shared across diverse contexts. More importantly, they typically suppress biased outputs without explicitly modeling the statistical dependence between model outputs and the underlying stereotype concepts. We propose Acmite, a lightweight concept-guided framework for targeted and selective debiasing. Acmite represents stereotypes as structured semantic concepts and uses maximal marginal relevance (MMR) to select diverse concepts for debiasing. Inspired by mutual information minimization, it approximates this dependence with token-level KL divergence while preserving task semantics. A lightweight LoRA adapter is trained with the base model frozen and activated at inference time only when the input is sufficiently similar to stereotype-related concepts; otherwise, the original model is used directly. We evaluate Acmite on BBQ, CrowS-Pairs, and StereoSet, and assess general capability preservation on ARC-Challenge, GSM8K, and PIQA. Experiments across three LLMs show that Acmite effectively mitigates gender bias across complementary evaluation formats while maintaining competitive performance on bias-unrelated tasks. Anonymous code and data are available at https://anonymous.4open.science/r/Acmite-18E2/.

cs.CL↗
arXiv2026

Belief-Aware Multi-Agent Path Finding under Map Uncertainty

Multi-Agent Path Finding (MAPF) aims to find collision-free paths for multiple agents in a shared environment. Classical MAPF assumes that all static obstacles are known in advance, but real-world environments can change unexpectedly due to fallen objects, spills, or other local disturbances. When such changes are spatially correlated, an observation can inform traversability estimates beyond the observed location. Prior approaches address uncertainty in traversability through contingent plans or replanning based on direct observations, but do not leverage this spatial dependence to infer the traversability of nearby unobserved locations. As a result, they cannot use one observation to anticipate nearby unobserved obstacles that may cause costly rerouting later. We focus on Belief-Aware MAPF, where map discrepancies are fixed during execution but initially unknown, and observations can be informative beyond the observed location. We propose Multi-Agent Gaussian belief Inference for Coordination (MAGIC), a framework that updates a shared belief about traversability online based on agents' observations. MAGIC uses a Gaussian Markov Random Field and Gaussian Belief Propagation to approximately infer traversability and construct detour-aware costs for standard MAPF planners. Our experiments on MAPF benchmarks show that MAGIC reduces the executed sum of costs compared to existing approaches on 96.3% of instances, across several planner families and teams of up to 800 agents, demonstrating its applicability to large-scale MAPF problems.

cs.AI↗
arXiv2026

PACE: Policy-Native Adaptive Decision Timing for Long-Horizon Reasoning

Long-horizon reasoning requires deciding not only what actions to take, but how many to execute open-loop before replanning. This number, the execution depth, balances replanning cost against compounding execution errors. Most current systems either fix the execution depth as a hand-tuned scalar or adjust it at inference time with heuristic rules decoupled from the policy; we argue both can be suboptimal. In this work, we treat the execution depth as a learnable, history-conditioned variable of the policy itself, and propose PACE, a model-native VLM policy that jointly predicts what to execute and for how long under a hard decision budget. We evaluate PACE on four long-horizon environments spanning fully and partially observable settings and visual and textual modalities: Sliding Puzzle, Sokoban, ALFWorld, and ScienceWorld. Across all settings, PACE improves success rate by 3.1 to 15.6 percentage points, while using fewer decisions on average, achieving empirical success-decision Pareto dominance.

cs.AI↗
arXiv2026

Bandwidth-Aware and Cost-Efficient Pipeline Parallel Scheduling in Geo-Distributed LLM Training

The rapid evolution of large language models (LLMs) has made geographically distributed training necessary due to GPU scarcity within a single cloud region. In such cross-region settings, Pipeline Parallelism (PP) is communication-efficient, yet scheduling PP remains challenging under heterogeneous inter-region bandwidth and regional electricity prices. Existing schedulers are either delay-first, incurring high electricity cost, or cost-first, relying on rigid resource allocation that prolongs Job Completion Time (JCT). They are also ineffective at optimizing execution order in multi-tenant environments, where long-running and bandwidth-intensive jobs can cause head-of-line (HoL) blocking and degrade overall performance. To this end, we propose BACE-Pipe, a bandwidth-aware and cost-efficient pipeline scheduling framework for LLM training across geo-distributed clusters. BACE-Pipe first introduces a dynamic job prioritization mechanism that optimizes execution order by jointly considering job characteristics (e.g., computation time) and real-time network utilization. It then employs a bandwidth-aware pathfinder to identify feasible cross-region pipeline paths that satisfy communication constraints, thereby preventing communication from stalling the pipeline. Among all feasible paths, a cost-minimizing allocator determines the optimal GPU placement strategy by preferentially assigning resources to regions with lower electricity prices. Consequently, BACE-Pipe mitigates HoL blocking, improves resource utilization, and simultaneously reduces both JCT and total electricity cost. Extensive simulations show that BACE-Pipe reduces average JCT by 27.9%--64.7% and total electricity cost by 12.6%--30.6% compared with state-of-the-art baselines.

cs.DC↗
arXiv2026

Decoupling Intention from Trajectory: A Representational Deduction Framework for World Action Models

World Action Models (WAMs) aim to construct a unified architecture capable of understanding world state evolution and guiding to generative motion planning. However, existing visual branches focus on predicting static visual observation, rather than reflecting potential transition information that captures the evolution of world states under motion interactions. This leads to representational entanglement between high-level physical condition evolution and low-level action trajectory generation within the Action Model, creating a structural bottleneck while weakening the predictive capability of world evolution modeling for action generation. We propose PILOT (Physical Inference for Latent Optimized Trajectories), whose core Representational Deduction (RD) bridges this gap by integrating motion thought-of-chain (CoT) guidance as a native model capability. Specifically, RD aims to encourage the action branch to explicitly model potential state transition tokens, which are retained as CoT in the reasoning space to guide fine-grained motion trajectory. Experiments demonstrate that RD not only significantly improves the success rate and generalization ability of WAMs in complex robotic manipulation tasks but also enhances the model's physical interpretability by decoupling high-level motion semantics from low-level trajectory details. Furthermore, the abundant state transition supervision signals introduced by RD effectively alleviate the sparse supervision in action generation, enabling it to serve as an efficient few-shot real-robot fine-tuning strategy and demonstrating superior scalability for migration to mainstream WAM architectures.

cs.RO↗
arXiv2026

From Visual Search to Movement Control: A Priority Field for Artificial Agents

Human spatial attention is widely conceptualized as being guided by a priority map that integrates perceptual salience, current goals, and past experiences. Here, we extend priority-based computation to movement control in artificial agents. We first introduce a lightweight model of visual search based on an integrated priority map. Trained on human saccades, it reproduced key behavioral patterns, including oculomotor suppression and history-driven selection. Extending the search model, we equipped an artificial agent with a priority field and evaluated its performance in a reach-avoid task that required reaching a goal destination while avoiding moving obstacles. Compared with alternative architectures, priority-field agents trained more efficiently and performed better in unseen, complex scenarios, even from simple demonstrations. Adding a simple memory mechanism also produced human-like, history-driven effects in anticipating the likely location of the upcoming goal. These findings suggest that priority-based computation may provide a promising foundation for movement control in artificial agents.

cs.RO↗
arXiv2026

From Detection to Understanding: TAR and TAR-Bench for Multi-Task Traffic Anomaly Reasoning

Detecting a traffic anomaly does not establish whether a video-language model can explain what happened, localize it in time, or identify its causes. We introduce TAR (Traffic Anomaly Reasoning) and TAR-Bench, paired resources for training and evaluating these complementary capabilities across 10 tasks spanning question answering, temporal reasoning, and scene understanding. TAR contains 44,040 automatically generated annotations with chain-of-thought traces for 3,670 CCTV videos from eight public datasets. TAR-Bench provides 960 human-curated annotations for 80 held-out clips from 17 public YouTube videos. Evaluation of eleven vision-language models reveals a gap between question-answering performance and temporal or scene reasoning. Progressively adding task groups during supervised fine-tuning improves aggregate performance on both Cosmos-Reason2-8B and Qwen3-VL-8B-Instruct. Training on all 10 tasks raises their mean benchmark scores from 34.3 to 55.7 and from 30.9 to 53.9, respectively. These results support joint training across complementary tasks as a promising approach to traffic anomaly understanding, while highlighting persistent limitations in temporal precision and causal attribution. TAR and TAR-Bench serve as the official training and in-domain evaluation resources for AI City Challenge 2026 Track 3.The dataset is available at https://huggingface.co/datasets/nvidia/PhysicalAI-Traffic-Anomaly-Reasoning

cs.CV↗
arXiv2026

Polarization and correlation effects in $γγ\to W^-W^+$ at Next-to-Leading-Order Electroweak Accuracy

We study the next-to-leading order (NLO) electroweak (EW) corrections to the polarization and correlation of $W$ boson pairs produced through photon-photon fusion in proton-proton ultra-peripheral collisions at the LHC, with both $W$ bosons decaying leptonically. The azimuthal correlation distributions of the charged leptons, which receive no contribution from anomalous quartic gauge couplings (AQGCs) at linear order, are perturbatively stable at the 0.40\% level, whereas individual correlation coefficient $C_{2,2}$ receives relative corrections of up to $-33.84\%$. Some of correlation coefficients that vanish at leading order (LO) in the Standard Model but are sensitive to AQGCs acquire non-zero one-loop values of order $10^{-4}-10^{-3}$. We quantify the region of the ($a_0/Λ^2,a_c/Λ^2$) plane in which these loop-induced baselines compete with the AQGC contributions and find that, while irrelevant at the current experimental sensitivity, they must be included once the correlation coefficients are measured at the per-mille level.

hep-ph↗
arXiv2026

Capacity Analysis and Joint Gaussian Beam Pattern Optimization for Positioning-Assisted Communications

Positioning-assisted beamforming is a novel enabling technology for massive multiple-input multiple-output in modern wireless communications, as it provides reliable beam steering without requiring explicit channel state information. However, the channel capacity analysis under positioning error remains mathematically intractable, which limits both performance characterization and beamforming design. In this work, we derive instantaneous and ergodic channel capacity approximations in closed forms for both two-dimensional and three-dimensional positioning-assisted beamforming systems. Based on the new expressions, we provide closed-form optimal joint Gaussian beam pattern that maximizes the asymptotic ergodic capacity. Numerical results verify the theoretical capacity expressions and the optimal beam pattern. The derived expressions enable efficient beam design to maximize channel capacity and provide a theoretical basis for positioning-assisted beamforming design.

cs.IT↗
arXiv2026

RPMem: Learning Long-Term Recurrent Parametric Memory Across Sessions for LLM Agents

Long-running LLM agents require memory that persists and evolves across sessions. Text-based memory retrieves and reconstructs past interactions at every query, making long-horizon performance increasingly dependent on retrieval quality and contextual reasoning as histories grow. Parametric memory encodes experience directly into model computation, but existing approaches provide limited support for cross-session memory evolution. Their coupling to a specific backbone further restricts memory reuse after model replacement. We introduce RPMem, a two-stage architecture that compiles each session into a model-independent latent memory through forward computation and selectively integrates it with retained memory via a task-trained recurrent gate. The consolidated memory is then mapped to backbone-specific low-rank adaptation (LoRA) parameters, allowing the encoding capability to transfer when the backbone is replaced. Evaluation across three long-term memory benchmarks and five diverse backbones demonstrates broad generalization with near-constant update cost and memory footprint. With Qwen3-8B on PERMA, RPMem reaches 85.52%, outperforming the strongest parametric and text-based baselines by 5.32 and 12.98 percentage points, respectively. Ablations validate the complementary roles of session compilation and cross-session consolidation, while dynamics analyses reveal that the gate acquires task-specific memory integration strategies. These results establish RPMem as a lifecycle-independent parametric memory framework that maintains evolving cross-session memory that remains reusable across backbone replacements. Our implementation is available at https://github.com/Quark-Medical/rpmem/tree/main.

cs.CL↗
arXiv2026

ACLArena: Agent Continue Learning in Multi-stage Post-training

Building general-purpose agents for industrial deployment requires integrating multiple capabilities, each typically acquired at a distinct stage of training. Yet there is currently no well-established recipe for Agent Continual Learning (ACL), with little understanding of the trade-offs among existing integration paradigms. To address this gap, we introduce ACLArena, a framework for comprehensively studying, analyzing, and evaluating ACL. We first build a sequential training pipeline and conduct an in-depth analysis that explains the mechanisms of forgetting and generalization from two complementary perspectives, the model level and the token level. Guided by these analyses, we systematically compare multi-teacher on-policy distillation, self-distilled fine-tuning, and model merging to assess their ability to recover previously learned capabilities while preserving newly acquired ones. Through extensive experiments, we develop a detailed understanding of how capabilities transfer across stages. Finally, we propose a new ACL recipe that combines offline replay over high-quality trajectories with a routed network of multiple LoRA experts each specialized via RL, substantially improving the agent's ability to learn across multiple domains. Comprehensive experiments on four reasoning and agentic tasks, evaluated under both in-domain and out-of-domain settings, demonstrate the value of our analysis and the effectiveness of our approach.

cs.AI↗
arXiv2026

Dengue fever model with impulsive intervention in a periodically varying environment

Pulse interventions represent a highly effective measure for infection control, as they influence disease transmission through short-term actions. In addition, the habitat ranges of dengue vectors and hosts exhibit periodic variations driven by environmental and climatic factors. To investigate the effects of impulsive interventions and domain evolution on disease transmission, we propose a dengue fever reaction-diffusion model that incorporates impulsive perturbations in a periodically varying domain. By applying the Poincar$\acute{e}$ map and the Krein-Rutman theorem, we establish the existence of the principal eigenvalue for the periodic eigenvalue problem with impulses, thereby extending previous studies on reaction-diffusion equations in fixed domains without impulsive effects. Sufficient conditions governing the long-term dynamics of periodic solutions are derived using the comparison principle and monotone iteration theory. Numerical simulations corroborate the theoretical findings and elucidate the effects of impulsive-intervention intensity and periodic domain evolution on disease transmission patterns. Our results indicate that increasing the intensity of pulse interventions suppresses disease transmission, whereas a larger magnitude of domain variation impedes disease control.

math.AP↗
arXiv2026

On a nonlocal dengue model with moving boundaries: The impact of dry-wet season and impulsive intervention

Seasonal succession and human intervention profoundly influence dengue fever transmission, and both humans and mosquito vectors can disperse over long distances. This paper proposes a nonlocal reaction-diffusion dengue model with moving boundaries that incorporates dry-wet seasonality and impulsive intervention. We first present the existence and uniqueness of the global solution to the model, then introduce the definition of the generalized principal eigenvalue, prove its uniqueness when the kernel functions coincide via the method of separation of variables, and further discuss its associated properties. Using this eigenvalue as an index, we obtain a spreading-vanishing dichotomy and derive two criteria that characterize the spreading and vanishing states, respectively. Our results extend both the nonlocal diffusion model without impulsive intervention and the model with impulsive intervention but without seasonal succession. Finally, we validate our theoretical findings and investigate how various factors affect dengue transmission through numerical simulations. The numerical results indicate that a shorter dry season makes disease containment more difficult, whereas more intensive impulsive intervention enhances control efficacy.

math.AP↗
arXiv2026

DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

The widespread adoption of long-horizon agents has made model workloads increasingly input-heavy. Although prior work has substantially reduced the cost of long-context computation, prefill remains computationally expensive, and large KV caches continue to strain HBM and SSD capacity and data-transfer bandwidth. Together, these compute, storage, and bandwidth demands constitute the primary bottleneck to further lowering deployment costs. To address this challenge, we introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. With its Causal Encoder-Decoder (CED) architecture, the model activates 16B parameters per token during decode but only 8B parameters during prefill, substantially improving cost efficiency for agentic workloads. To push the limits of KV cache compression, DeepSeek-V4.1-Flash combines cross-layer KV cache reuse in Compressed Sparse Attention 2 (CSA2) with FP4 KV caching. These designs reduce its global KV cache footprint (always in HBM) to 890 bytes per token, roughly 1/4 of the corresponding footprint of DeepSeek-V4-Flash. Further, through a dedicated deployment optimization known as SWA Bounded Replay, DeepSeek-V4.1-Flash reduces its persistent KV cache footprint (always on SSD or in host memory) to roughly 1/8 of that of DeepSeek-V4-Flash. Despite its much smaller KV cache footprint, the model delivers substantially better performance than the baseline. In addition, we streamline the DeepSeek-V4 architecture and introduce several efficient architectural extensions. We pretrain DeepSeek-V4.1-Flash on a multimodal corpus comprising 45T tokens and conduct comprehensive post-training, yielding strong performance across diverse text-based and multimodal agentic scenarios. Model checkpoints are available at https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash.

cs.CL↗
arXiv2026

ThinkFlow: Self-Evolving Probabilistic Latent Memory for Lifelong Conversational Agents

Lifelong conversational agents rely on memory systems to maintain deep, context-aware interactions with users. However, existing explicit textual memory pipelines suffer from a severe information bottleneck, often losing subtle behavioral patterns and emotional shifts. Furthermore, being typically static post-deployment, they cannot autonomously adapt to personal habits and preferences without manual feedback. Cognitive science, however, suggests that humans maintain mental models purely in a latent space and continuously refine them through predictive coding. Inspired by this, we propose \textbf{ThinkFlow}, a novel end-to-end latent memory framework for lifelong conversational agents. ThinkFlow bypasses the text bottleneck by dynamically compressing conversational flows into probabilistic latent memory skills, autonomously consolidating complex user states into disentangled, continuous vectors without semantic interference. To break this barrier, we introduce a test-time evolution paradigm. By coupling teacher-guided latent alignment to bootstrap the initial state with a self-supervised next-user-utterance prediction task for continuous refinement, the framework successfully overcomes cold-start challenges and achieves label-free lifelong personalization. Extensive experiments on long-term conversation benchmarks demonstrate that ThinkFlow significantly outperforms prevailing memory systems, providing highly personalized and contextually accurate responses over extended multi-session interactions.

cs.AI↗
arXiv2026

Interactive Memory Learning for Long-Term Conversations

Recent advancements in large language models have significantly enhanced the capabilities of agents in modeling long-term conversations. Despite these successes, existing approaches typically adopt a static heuristic paradigm, where information is passively archived without adaptive memory valuation. Consequently, these methods fail to self-evolve or align their memory management with evolving user needs. To address this, we propose ICML (InteraCtive Memory Learning), a multi-agent framework that transforms the memory mechanism from a passive archive into a learnable, interactive memory policy. Specifically, we first employ a session synthesis pipeline to generate expert data, facilitating rapid test-time adaptation in unseen scenarios. Building on this, ICML utilizes an online reinforcement learning mechanism where a Planner agent selectively encodes high-value information and a Trigger agent dynamically retrieves it to optimize response quality, whereby the two agents co-evolve through continuous interaction feedback. Crucially, both agents are synchronized through a delayed reward mechanism that propagates future feedback back to earlier storage decisions, ensuring memory policies are precisely aligned with user expectations. Experimental results demonstrate that ICML significantly outperforms strong baselines, exhibiting the unique capability to continuously improve response quality as interactions accumulate.

cs.AI↗
arXiv2026

ITA-LaCAM: A Complete and Scalable TAPF Solver via Assignment-Aware Configuration-Space Search

Combined Target Assignment and Path Finding (TAPF) requires assigning targets for agents while simultaneously planning collision-free paths. We present ITA-LaCAM, a complete and scalable TAPF solver inspired by LaCAM and ITA-CBS. In ITA-LaCAM, each joint-configuration node carries an agent-to-target matching. When a successor is generated, ITA-LaCAM incrementally repairs the matching for the agents that moved and uses the targets to guide PIBT successor generation. This design enables adaptive reassignment without explicitly enumerating the combinatorial assignment space, while preserving LaCAM's completeness and scalability. Across 9,760 benchmark instances on eight maps with 5--200 agents, ITA-LaCAM solved 100% of the instances, compared with 95.6% for IR-TAPF configured with DBS-Hungarian. ITA-LaCAM found an initial solution faster in 84.0% of the comparisons and achieved a lower sum of costs in 65.0% of the instances solved by both methods.

cs.RO↗
arXiv2026

Fundamental Limits of Joint Target Detection and Parameter Estimation - Characterizing Mixed-State Sensing Limits via Posterior Entropy Volume

The development of integrated sensing and communication calls for a unified theoretical foundation for sensing. This paper models target-presence patterns and continuous physical parameters as a mixed discrete-continuous state $Ξ$ on a branched reference measure, and treats posterior entropy volume and joint mutual information as two complementary representations of the same limit. Entropy volume carries physical units, can be compared with engineering scales such as resolution cells, and remains meaningful when the number of active targets varies; mutual information is dimensionless, invariant to coordinates and units, and additive through the chain rule. We define entropy number, entropy volume, and mixed entropy volume for discrete, continuous, and mixed states, respectively. The maximum-entropy principle for the uniform distribution on a support of fixed measure explains the measure-theoretic meaning of the exponential entropy scale, while the main limit follows from a mixed asymptotic equipartition property and a posterior probability-volume inequality through conditional typical sets. For asymptotically reliable high-probability sensing regions, the minimum achievable first-order posterior mixed entropy volume equals the prior mixed entropy volume multiplied by $2^{-I(Ξ;Y)}$, where $I(Ξ;Y)=I(V;Y)+I(X_V;Y\mid V)$. Thus, detection and estimation contributions multiply in the volume domain and add in the bit domain, and one sensing bit halves the posterior effective measure. We further prove that posterior-preserving cascades attain the direct-inference limit, while arbitrary intermediate compression incurs the exact information loss $I(Ξ;Y\mid Z)$. Numerical results for a single-target presence-range model illustrate the information composition, posterior entropy-volume contraction, and cascade-interface loss.

cs.IT↗