arXiv subjects

Jiayu Chen

Publications and source records attributed to Jiayu Chen.

At least 19 recordsLinked to original sources

arXiv2026

Hierarchical Deep Counterfactual Regret Minimization

Imperfect Information Games (IIGs) are used to model games under uncertainty or lack complete information. Counterfactual Regret Minimization (CFR) is one of the most successful families of algorithms for IIGs. The integration of skill-based strategy learning with CFR could potentially mirror more human-like decision-making and improve learning on complex IIGs. It enables the learning of a hierarchical strategy, wherein low-level components represent skills for solving subgames and the high-level component manages the transition between skills. In this paper, we introduce the first hierarchical version of Deep CFR (HDCFR), an innovative method that boosts learning efficiency in tasks involving extensively large state spaces and deep game trees. Notably, HDCFR enables learning with predefined (human) expertise and extracting skills transferable to similar tasks. We first present the algorithm and establish its theory in a tabular setting, including hierarchical CFR update rules and a variance-reduced Monte Carlo sampling extension for the model-free setting, where backtracking is infeasible. We then extend HDCFR to large-scale tasks via deep learning objectives that match the tabular targets under exact function fitting. Code: https://anonymous.4open.science/r/HDCFR_RUN-677B.

cs.LG↗

arXiv2026

BigMoMo: Efficient Inference of Large-Scale MoE with Speculative Decoding on Mobile Devices

Mixture-of-Experts (MoE) models expand language model capacity on smartphones, but expert offloading remains constrained by limited DRAM capacity and costly data movement. Sequential token routing couples expert execution to fragmented flash reads and multistage NPU preparation, leaving sparse computation stalled on weight transfers. Each transfer serves few tokens before execution moves on. We exploit the multi-token verification window of speculative decoding to decouple expert movement from single-token execution, enabling weight reuse, contiguous flash reads, and load-compute overlap. We present \textsc{BigMoMo}, a mobile MoE runtime that exploits this window across the memory hierarchy. It prunes speculative branches and expert activations using acceptance rates, routing impact, and movement cost; reorganizes on-flash experts according to runtime co-loading patterns; and batches ready experts to overlap NPU computation with pending transfers. Across four MoE models and five benchmarks on two mobile platforms, \textsc{BigMoMo} achieves mean decoding speedups of $4.83\times$ over on-demand autoregressive offloading and $1.82\times$ over the best speculative MoE baseline, supporting MoE models up to 30B parameter.

cs.AR↗

arXiv2026

Stable-MM-R1: Anchoring Multimodal Reasoning Dynamics via Entropy-Guided Stratification

While Reinforcement Learning (RL) effectively incentivizes reasoning in Large Language Models, current pipelines are hindered by training instability and rapid entropy collapse. These limitations often stem from "Rollout Silencing" and low-quality gradient signals in standard sampling procedures. In this work, we propose a robust, data-centric framework to stabilize RL training. We first introduce Potential-Aware Query Mining (PAQM), which filters data dynamically to focus on the "Distillation Zone"---samples with high potential for capability elicitation. Furthermore, we present Hybrid Stratified Replay (HSR), a novel mechanism that restructures batches by stratifying rollouts based on Path Entropy, a rollout-level confidence proxy, and outcome reward. Within each optimization step, HSR reuses current-policy "Stability Anchors" and "Hard Negatives" to construct high-contrast optimization groups, then clears its buffers before the next step. This approach mitigates entropy collapse while improving the utilization of learning signals under limited compute. Our method outperforms strong baselines on complex reasoning tasks, offering a principled solution for stable and efficient RL fine-tuning.

cs.LG↗

arXiv2026

Continual Policy Consolidation for Lifelong Robot Learning

Building a generalist robot policy requires continuously integrating new skills while preserving previously acquired behaviors. Directly optimizing a single policy over a growing task stream is difficult because robotic interaction is expensive, task distributions are heterogeneous, and sequential updates induce interference. To address these problems, we propose continual policy consolidation (CPC), a teacher--student framework that combines continual policy distillation with prioritized experience replay and expandable experts. This architecture separates skill acquisition from policy consolidation: teachers are trained independently through reinforcement learning, and their behaviors are continually distilled into a central generalist student. This decomposition retains the practical strength of reinforcement learning for task-specialized training while casting student-side consolidation as a supervised policy-learning problem. To balance stability and plasticity as the task stream grows, the student combines an expandable Transformer-based mixture-of-experts architecture with prioritized trajectory replay. Extensive experiments show that the student recovers a large proportion of teacher performance while achieving near-zero forgetting. These results demonstrate a scalable route for consolidating independently acquired robot skills into a continually growing generalist policy.

cs.LG↗

arXiv2026

PAC-CF: Calibrating Irreversible Frontier Pruning in LLM-Guided Search

LLM-guided search is usually adopted to solve complex tasks by ranking and pruning top-$K$ candidates based on evaluator scores. However, irreducible bias still exists even if popular methods, such as repeated sampling, are applied to reduce variance. Consequently, pruning may remove every continuation that can reach a valid solution. In this paper, we propose Probably Approximately Correct Conformal Filtering (PAC-CF), which formulates tree pruning as a PAC-guaranteed decision problem. Theoretical analysis establishes how irreducible bias reduces the score separation for certified elimination. Native-Trace path calibration derives a conformal margin from the score deficit of verifier-valid continuations on held-out Native traces. During deployment, PAC-CF uses this calibrated margin in a direct score-gap filtering rule. Across diverse domains and state-of-the-art controllers, PAC-CF improves utility at various budgets while reducing all five measured workload metrics. Especially on pruning-aware ToolTree under a 100-request budget, replacing native top-$K$ improves equal-domain utility by 4.38 points while reducing physical requests by $18.95\%$ with a $23.76\%$ token reduction.

cs.LG↗

arXiv2026

Unified Condition-Action Modeling for Accurate One-Step Action Generation

Robot manipulation requires policies that are both accurate and efficient, as robot control must respond to changing observations under tight latency constraints. Recent diffusion and flow policies are promising, but they often treat conditions as auxiliary signals rather than jointly evolving them with action trajectories. We find that this limitation can be effectively mitigated by a \textbf{simple yet effective unified condition-action modeling design} that represents conditions and actions in a shared token space, allowing a compact model to achieve high performance while improving both inference speed and accuracy. Therefore, we propose UCA-Flow, a unified condition-action modeling framework for accurate one-step action generation. Our method unifies observation conditions, timestep conditions, interval conditions, and action tokens into a single sequence, and processes them with a Unified Condition-Action Transformer for joint condition-action representation learning. As a result, condition representations are dynamically reconstructed according to the current generation stage, highlighting information most relevant for action refinement. Furthermore, we introduce an improved dual-pass supervision scheme over $u$ and $v$ for stronger optimization of unified condition-action modeling. UCA-Flow improves the average success rate by 9.3 percentage points over the strongest baseline, while achieving $45.6\times$ and $33.4\times$ speedups over DP3 and Simple DP3, and remaining $4.3\times$ and $2.3\times$ faster than one-step FlowPolicy and MP1, respectively.Project page: https://uca-policy.github.io/UCA.github.io/.

cs.RO↗

arXiv2026

An Evidence-Grounded Multi-Agent System for High-Level Bio-Robot Design

In this paper, a bio-robot is an engineered living or biohybrid system in which living cells perform one or more core functions, such as sensing, information processing, actuation or output. We focus on systems whose cell-based functions are programmed by genetic circuits; physical movement is optional. Designing such a system requires translating application requirements into sensing, logic or memory, output, assembly, host and containment modules, while grounding each choice in traceable parts and evidence. We present micro_biorobot_agent, an offline multi-agent system built on Qwen3.5-27B. The system combines requirement analysis, module-specific retrieval, candidate assembly, conflict checking, local repair, independent review and validation over an integrated library of 23,762 records covering biological parts, measured combinations, literature-supported relationships and actuation evidence. Deterministic output checks align the final report with the retrieved part set and correct false gaps, unsupported part mentions and source-tracking errors. On two author-developed evaluation sets of 50 queries each, the system obtains mean overall scores of 7.35 and 8.04, the highest among the seven evaluated systems; on Scenario Design it exceeds the runner-up by 2.23 points. A 50-query paired ablation shows that the source-tracking check reduces false-gap incidents from 15 to 3, an 80% reduction, and increases source accuracy by 0.75 points. This paper reports the Qwen3.5-based v1 system and evaluates high-level design reports rather than experimentally validated circuits.

cs.MA↗

arXiv2026

EchoCache: Energy-Guided Cross-Modal Caching for Efficient Audio-Driven Video Generation

Audio-driven video generation (A2V) has achieved promising progress in synthesizing temporally coherent and audio-visually aligned videos, yet its inference remains expensive due to the iterative denoising process of diffusion models. Existing caching methods mainly exploit temporal redundancy in visual features while overlooking the cross-modal alignment of A2V, where audio drives visual generation with highly non-uniform temporal importance. In this paper, we identify two levels of misalignment in existing A2V caching methods: temporal-semantic and computation-storage misalignment. To address them, we propose EchoCache, an energy-guided cross-modal caching framework for efficient A2V generation. EchoCache leverages audio time-frequency energy as a saliency anchor to guide latent-level cache updates and further introduces a dynamic timestep-latent caching mechanism with quantized cache management for joint efficiency and memory optimization. Extensive experiments on mainstream A2V models show that EchoCache consistently improves the latency-quality trade-off while preserving generation quality and audio-visual consistency. In particular, on Wan2.2-S2V over the EMTD benchmark, EchoCache achieves a 2.46x speedup with the best overall performance. Code is available at https://github.com/IF-LAB-PKU/EchoCache.

cs.CV↗

arXiv2026

Traj-LeWM: Path-Aware World-Model Planning via Latent Trajectory Cost

LeWM is a lightweight visual world model that learns latent dynamics end-to-end from pixels and ranks candidate action sequences by the distance between their predicted endpoints and the goal. However, LeWM has two limitations. First, during training, it learns local next-step transitions without evaluating complete trajectories relative to the task goal. Second, during planning, it ranks candidates solely by predicted endpoint distance. Because model predictions may differ from actual execution outcomes, the candidate whose predicted endpoint is closest to the goal may not perform best when executed in the environment. The evolution of the complete predicted trajectory can therefore provide complementary information beyond endpoint distance. To address these limitations, we propose Traj-LeWM, which retains LeWM's local-dynamics objective and endpoint score while introducing a goal-conditioned latent trajectory cost (LTC) that aggregates trajectory-level information as a complementary signal. During training, LTC-based trajectory-preference supervision complements next-step prediction in shaping the shared representation. During planning, LTC is combined with endpoint distance to incorporate intermediate-path information into candidate ranking. With joint endpoint-plus-LTC scoring, Traj-LeWM outperforms LeWM on Push-T, OGBench-Cube, Reacher, and Two-Room by $3$, $14$, $7$, and $7$ percentage points, respectively. Controlled experiments and ablations further verify the complementary roles of trajectory-level representation shaping and path-aware candidate ranking.

cs.AI↗

arXiv2026

AgilePE: Autonomous UAV Pursuit-Evasion via Self-Play Reinforcement Learning

Autonomous pursuit-evasion is a fundamental challenge for Unmanned Aerial Vehicles (UAVs), requiring rapid decision-making under tightly coupled dynamics and continuously changing opponent behaviors. Traditional rule-based or differential-game approaches often struggle with high-dimensional aerial interactions and agile maneuvering. We present AgilePE, a complete system for autonomous UAV pursuit-evasion via self-play reinforcement learning. AgilePE integrates agile low-level control, competitive policy optimization, and sim-to-real deployment in a unified framework. The policy directly maps onboard state observations to Collective Thrust and Body Rates (CTBR) commands, enabling end-to-end agile maneuvering without intermediate trajectory planners or waypoint controllers. For training, we use competitive self-play with Prioritized Fictitious Self-Play (PFSP) and a diversified opponent pool, enabling agents to improve against historical policies while stabilizing optimization and reducing policy oscillation. This process leads to the emergence of sophisticated pursuit and evasion strategies. For real-world deployment, we develop a hardware-aligned simulation pipeline that models actuator-response dynamics, communication latency, and domain randomization. The learned policies transfer zero-shot to real quadrotors without task-specific tuning. Real-world experiments reproduce pursuit-evasion tactics observed in simulation, including rapid dodging and flanking, and demonstrate interactive two-agent zero-shot deployment.

cs.RO↗

arXiv2026

EnergyBridge: Benchmarking Household Energy Management, User Participation, and Grid Flexibility

Residential virtual power plants (VPPs) can provide grid flexibility by shifting household demand, but physical flexibility becomes dependable capacity only when residents authorize a plan and the promised response is delivered. Existing benchmarks evaluate control but omit event-specific authorization. We present EnergyBridge, a benchmark and agent framework connecting capacity reporting, household authorization, and physical execution. It combines region-specific EnergyPlus environments for Tianjin and Berlin with an LLM-based User Participation Simulator. Against 584 persona- and event-matched human role-play judgments, the LLM-based User Participation Simulator preserves method ordering with a 5.3-point mean absolute acceptance error. Across conventional controllers and agent baselines, EnergyBridge achieves the highest simulated authorization, lowest event-window energy, and the most reliable capacity commitment in both regions. We release human data and codes for reproducible human-centered grid-flexibility research: https://github.com/Agentic-Intelligence-Lab/EnergyBridge.

cs.AI↗

arXiv2026

Can Vision-Language-Action Models Learn from Real-World Data Continually without Forgetting?

Vision-Language-Action (VLA) models provide a promising foundation for general-purpose robotics, yet their real-world deployment demands the ability to continually acquire new skills without forgetting prior ones. While recent studies have explored continual learning for VLA models in simulated settings, the challenge remains largely unexamined under realistic physical conditions. To bridge this gap, we construct a real-world continual learning benchmark comprising ten diverse sequential manipulation tasks across both single-arm and bimanual configurations. Through extensive experiments on this benchmark, we find that naive sequential fine-tuning leads to severe catastrophic forgetting, whereas a well-configured experience replay (ER) approach can effectively mitigate forgetting and outperform joint multi-task training under equivalent computational budgets. Notably, by synthesizing our empirical findings, we successfully achieve stable continual learning across the full 10-task heterogeneous stream, retaining previously acquired capabilities while adapting to diverse new skills in real-world deployment. This work presents an empirical study grounded in real-world continual VLA learning and offers actionable insights for deploying robust, long-lived robotic policies.

cs.RO↗

arXiv2026

MoECa: Aligning Feature Reuse with Expert Decomposition in Diffusion Transformers

Diffusion Transformers with Mixture-of-Experts (DiT-MoE) improve model capacity under sparse activation, but diffusion inference is still bottlenecked by redundant computation across timesteps. Existing caching methods mainly operate at the token level, which becomes suboptimal in DiT-MoE because each token update is internally decomposed into multiple routed expert branches. Our analysis shows that cross-timestep redundancy in DiT-MoE is better characterized at the expert-branch level than at the whole-token level. Based on this observation, we propose MoECa, a fine-grained caching framework that performs branch-level feature reuse across timesteps. MoECa further introduces expert-aware adaptive control and synchronized cache updates across MoE and attention paths to maintain stable intermediate states. Experiments on multiple DiT-MoE models show a favorable speed--quality trade-off, with up to 2.93$\times$ speedups while preserving generation quality.

cs.LG↗

arXiv2026

When and Where to Look: Adaptive Visual Evidence Scheduling for Efficient Long Video Understanding

Efficient long-video understanding requires vision--language models (VLMs) to reason over a small number of frames selected as sparse visual evidence. Existing relevance-based methods rely on static one-shot selection with fixed frame budgets and candidate pools, while agent-based schedulers achieve adaptivity through costly multi-round reasoning and interactive search. We propose EcoFrame, a training-free framework for low-overhead query-adaptive visual evidence scheduling. EcoFrame leverages the VLM's inference feedback to determine when to increase the frame budget and where to search for additional candidate evidence. Specifically, entropy-gated budget scheduling uses output uncertainty to stop early when the current evidence is sufficient or progressively expand the frame budget otherwise. Meanwhile, attention-guided candidate proposal converts frame-level attention into a temporal prior, enabling dense local search in informative regions while preserving global coverage when attention is diffuse. Experiments on Video-MME, LongVideoBench, and MLVU demonstrate that EcoFrame achieves a better accuracy--efficiency trade-off across multiple VLM backbones. On Qwen2.5-VL, EcoFrame achieves an average accuracy of 64.4, surpassing BOLT at 63.5, while providing a $1.85\times$ speedup over AKS and BOLT. Compared with the agent-based A.I.R., EcoFrame maintains comparable accuracy with up to a $13.5\times$ inference speedup. Code will be available at https://github.com/AK-DREAM/EcoFrame.

cs.CV↗

arXiv2026

Distributionally Robust Listwise Preference Optimization

Existing robust preference optimization for language-model alignment mainly studies pairwise supervision and places robustness at the dataset, prompt, or preference-pair level. We instead study listwise preference optimization under ranking-label uncertainty: given a prompt and a candidate list, the observed ranking over that list may be ambiguous due to annotator inconsistency, near-ties, lossy rankwise feedback, or reward-model noise. We propose a pointwise total-variation robust Plackett--Luce objective that directly robustifies the ranking label conditional on the candidate list. The robust loss admits an exact decomposition into the nominal PL loss plus a worst-case PL correction, and the worst-case ranking is obtained by sorting current implicit scores in ascending order, reducing the inner maximization from $K!$ enumeration to $O(K\log K)$. This tractable structure yields strong offline and online optimization guarantees. In the offline fixed-list setting, the robust objective is convex and projected stochastic subgradient reaches global $ε$-suboptimality with $O(ε^{-2})$ sample complexity. In the online policy-induced setting, where candidate lists are generated by the current policy, we establish weak convexity and $\widetilde O(ε^{-2})$ Moreau-envelope stationarity. Experiments in offline LLM alignment show that the proposed robust correction largely preserves performance under clean labels and improves robustness under noise. In online alignment, it makes reward-model-ranked candidate expansion more reliable and improves both reward-model and external GPT-4 judge metrics.

cs.AI↗

arXiv2026

Token Radius Attention for Efficient Video Generation

Video Diffusion Transformers (VDiTs) enable high-fidelity generation but incur quadratic cost from dense 3D self-attention. Existing head- and block-level sparse methods share computation budgets across queries, overlooking token-specific attention demand. We observe that retained density varies across queries yet correlates log-linearly with attention entropy, while dominant interactions form query-centered neighborhoods with token-dependent radii. Based on these findings, we propose Token Radius Attention (TRA), a training-free framework that maps query entropy to an analytic token budget and converts it into a temporally decayed radius without explicit key ranking. Fused entropy extraction, warm-up reuse, and block-sparse mask construction further reduce overhead. Across seven Wan2.1, Wan2.2, and HunyuanVideo T2V/I2V configurations, TRA retains only 9-19% of attention interactions and achieves 1.56x-2.05x speedup with competitive generation quality. Code is available at https://github.com/IF-LAB-PKU/Token-Radius-Attention.

cs.CV↗

arXiv2026

Bootstrap Policy Iteration for Stochastic Linear Quadratic Tracking with Multiplicative Noise

This paper studies the linear quadratic tracking problem for continuous-time stochastic systems with multiplicative noise. The proposed framework formulates the problem under an average cost criterion and separates the computation of the optimal feedback and feedforward gains. By developing a bootstrap policy iteration algorithm, we eliminate the restrictive a priori requirement for an initial mean square stabilizing feedback gain in existing policy iteration methods. Based on this iterative framework, an off-policy reinforcement learning algorithm is proposed to learn the optimal feedback gain directly from data. Using the learned feedback gain, the feedforward gain is subsequently obtained through a data-driven one-shot computation procedure. These components work together to provide a model-free solution to the stochastic optimal tracking control problem. The effectiveness of the proposed method is demonstrated through a numerical example.

eess.SY↗

arXiv2026

SAM3D-Guided Object-Centric Representation Alignment for Vision-Language-Action Models

Vision-Language-Action (VLA) models have shown strong potential for general robot manipulation, but most existing models rely on 2D visual-language backbones and lack fine-grained 3D understanding of target objects, especially under occlusion, pose variation, scale changes, and precise spatial interaction. We propose an object-centric 3D representation alignment framework built upon $π_0$, using SAM3D as a frozen 3D teacher to provide target-object 3D priors during training. Specifically, we localize task-relevant objects with object recognition models, generate corresponding object masks, and use SAM3D to extract dense object-level 3D representations, which are aligned with intermediate visual features of $π_0$. This enables the policy to internalize target-object 3D information while preserving the original RGB-language-to-action inference pipeline without requiring depth, point clouds, masks, SAM3D, or additional 3D modules at test time. Simulation experiments show consistent improvements, achieving 99.1\% on LIBERO and an average length of 4.11 on CALVIN. Real-world experiments further demonstrate that our method is particularly effective in long-horizon manipulation scenarios where the robot must focus on different target objects across multiple subtasks.

cs.RO↗