arXiv subjects

Chao Feng

Publications and source records attributed to Chao Feng.

At least 37 records · Page 2Linked to original sources

arXiv2026

Unified Motion-Action Modeling for Heterogeneous Robot Learning

We present Unified Motion-Action (UMA) Model, an approach that uses 3D object motion trajectories as a shared interface to bridge visuomotor control and dynamics modeling. UMA treats object motion and robot actions as co-evolving variables under a masked generative objective, in which the mask pattern determines both the supervision regime during pretraining and the inference mode at deployment. Using hindsight-relabeled motion contexts and a contrastive objective that disentangles task intent from scene geometry, UMA enables multi-task pretraining across heterogeneous data sources without requiring manually annotated task instructions. At deployment, the same pretrained parameters support motion-conditioned visuomotor control, motion-based dynamics modeling, and task adaptation from few-shot demonstrations. Pretrained on a mixture of robot demonstrations, human videos, and simulated data, UMA consistently outperforms state-of-the-art baselines specialized for each inference mode.

cs.RO↗

arXiv2026

AgentCanary: A Security Evaluation Framework for Autonomous AI Agents in Real Executable Environments

Autonomous AI agents have driven the transition from conversation to task execution, shifting security failures from textual deception to system compromise. Although security evaluation is crucial for proactive risk prevention, prior work is constrained by fundamental bottlenecks, including fragmented risk coverage, static or low-fidelity execution environments, and single-dimensional and coarse-grained assessment metrics. To address these challenges, we propose AgentCanary, a comprehensive security evaluation framework for autonomous AI agents. AgentCanary provides a systematic solution along three contributions. First, comprehensive risk coverage: we introduce an orthogonal Entry $\times$ Impact risk taxonomy that decouples how adversarial influence enters the agent from what harm it ultimately causes, and instantiate it as a scenario-aligned task suite spanning realistic deployment workflows. Second, a high-fidelity real executable environment: rather than static Q&A or mocked tool responses, agents interact with real tools against dynamically provisioned task artifacts, with persistent state across multi-step interactions that naturally supports long-horizon attack evaluation. Third, trajectory-grounded multi-dimensional evaluation: evaluation consumes the full agent trajectory rather than the reply text or a single tool call, enabling decomposed scoring along three orthogonal dimensions, Outcome Safety, Security Awareness, and Task Utility. We evaluate a broad set of frontier models on AgentCanary against multiple established adversarial attack methods across three agent frameworks. The results reveal that current agents often fail to recognize the attacks they face, particularly under compromised skills, persistent state, and long-horizon execution attacks, and provide a systematic baseline for developing more reliable and secure agent systems.

cs.CR↗

arXiv2026

Fully coherent short wavelength free-electron laser driven by a single sub-microjoule seed

High-repetition-rate, fully coherent extreme-ultraviolet (EUV) and X-ray free-electron lasers (FELs) are essential for advanced time-resolved ultrafast spectroscopies. While external seeding serves as the standard technique to achieve precise temporal coherence, conventional methods demand hundred-megawatt peak-power laser systems. Furthermore, advanced configurations like echo-enabled harmonic generation (EEHG) introduce the severe complexities of dual-laser synchronization. Together, these requirements fundamentally restrict operations to kilohertz repetition rates and compromise overall system stability. Here, we experimentally demonstrate a fully coherent EEHG-FEL driven by a single, sub-microjoule seed laser. By employing a direct-amplification enabled harmonic generation technique, we utilize an initial 0.4 microJ (2 MW peak power) ultraviolet seed to directly drive coherent lasing at nanometer wavelengths. By eliminating the need for extreme peak powers and multiple synchronized lasers, this approach significantly simplifies the seeding architecture and provides a practical and robust pathway toward megahertz-class, fully coherent EUV and X-ray light sources.

physics.acc-ph↗

arXiv2026

VGR: Visual Grounded Reasoning

In the field of multimodal chain-of-thought (CoT) reasoning, existing approaches predominantly rely on reasoning on pure language space, which inherently suffers from language bias and is largely confined to math or science domains. This narrow focus limits their ability to handle complex visual reasoning tasks that demand comprehensive understanding of image details. To address these limitations, this paper introduces VGR, a novel reasoning multimodal large language model (MLLM) with enhanced fine-grained visual perception capabilities. Unlike traditional MLLMs that answer the question or reasoning solely on the language space, our VGR first detects relevant regions that may help to solve problems, and then provides precise answers based on replayed image regions. To achieve this, we conduct a large-scale SFT dataset called VGR -SFT that contains reasoning data with mixed vision grounding and language deduction. The inference pipeline of VGR allows the model to choose bounding boxes for visual reference and a replay stage is introduced to integrates the corresponding regions into the reasoning process, enhancing multimodel comprehension. Experiments on the LLaVA-NeXT-7B baseline show that VGR achieves superior performance on multi-modal benchmarks requiring comprehensive image detail understanding. Compared to the baseline, VGR uses only 30\% of the image token count while delivering scores of +4.1 on MMStar, +7.1 on AI2D, and a +12.9 improvement on ChartQA.

cs.CV↗

arXiv2026

Phase-Stable Self-Modulation for GHz Continuous-Wave Ultrafast X-Ray Free-Electron Lasers

High-brightness femtosecond-to-attosecond pulses are indispensable for probing electron dynamics on their fundamental temporal scales. X-ray free-electron lasers (XFELs) at high repetition rates will facilitate high-statistics measurements and time-resolved studies that were previously inaccessible. Although energy recovery linacs (ERLs) are well suited for high-repetition-rate operation, their relatively low peak current poses a major challenge for generating intense ultrashort X-ray pulses. Here, we propose a completely laser-free scheme that fundamentally overcomes this bottleneck through a continuous, phase-stable self-modulation process. By interacting with its own coherently emitted terahertz radiation within a helical wiggler, the electron bunch naturally accumulates a robust, few-cycle energy modulation in its core, even when starting with the intrinsically low peak current typical of ERLs. A downstream dispersion chicane subsequently converts this energy modulation into an isolated, exceptionally sharp current spike. Start-to-end simulations based on a 1~GeV ERL light source demonstrate the feasibility of generating isolated soft X-ray pulses with an average peak power exceeding 4~GW and a pulse duration of about 1~fs at an unprecedented 1.3~GHz repetition rate. The proposed scheme offers a highly practical pathway for advancing ultrafast X-ray generation into the true continuous-wave regime, with transformative implications for the development of next-generation coherent light sources.

physics.acc-ph↗

arXiv2026

A Crowdsensing Intrusion Detection Dataset For Decentralized Federated Learning Models

This paper introduces a dataset and an experimental study on Decentralized Federated Learning (DFL) for Internet of Things (IoT) crowdsensing malware detection. The dataset comprises behavioral records from benign and eight malware attacks. A total of 21,582,484 original records were collected from system calls, file system activities, resource usage, kernel events, input/output events, and network records. These records were aggregated into 30-second windows, resulting in 342,106 data records used for model training and evaluation. Experiments on the DFL platform compare traditional Machine Learning (ML), Centralized Federated Learning (CFL), and DFL across different node counts, topologies, and data distributions. Results show that DFL maintains competitive performance while preserving data locality, outperforming CFL in most settings. This dataset provides a solid foundation for studying the security of IoT crowdsensing environments.

cs.CR↗

arXiv2026

TextPecker: Rewarding Structural Anomaly Quantification for Enhancing Visual Text Rendering

Visual Text Rendering (VTR) remains a critical challenge in text-to-image generation, where even advanced models frequently produce text with structural anomalies such as distortion, blurriness, and misalignment. However, we find that leading MLLMs and specialist OCR models largely fail to perceive these structural anomalies, creating a critical bottleneck for both VTR evaluation and RL-based optimization. As a result, even state-of-the-art generators (e.g., Seedream4.0, Qwen-Image) still struggle to render structurally faithful text. To address this, we propose TextPecker, a plug-and-play structural anomaly perceptive RL strategy that mitigates noisy reward signals and works with any textto-image generator. To enable this capability, we construct a recognition dataset with character-level structural-anomaly annotations and develop a stroke-editing synthesis engine to expand structural-error coverage. Experiments show that TextPecker consistently improves diverse text-to-image models; even on the well-optimized Qwen-Image, it significantly yields average gains of 4% in structural fidelity and 8.7% in semantic alignment for Chinese text rendering, establishing a new state-of-the-art in high-fidelity VTR. Our work fills a gap in VTR optimization, providing a foundational step towards reliable and structural faithful visual text generation.

cs.CV↗

arXiv2026

UAV Swarm Enabled Aerial Movable Antenna System for Low-Altitude Economy: From Far-Field to Near-Field Communication

Unmanned aerial vehicle (UAV) with the intrinsic three-dimensional (3D) mobility provides an ideal platform for implementing aerial movable antenna (AMA) system enabled by UAV swarm cooperation. Besides, AMA system is readily to achieve an extremely large-scale array aperture, rendering the conventional far-field uniform plane wave (UPW) model no longer valid for aerial-to-ground links. This paper studies the UAV swarm enabled near-field AMA communication, by taking into account the non-uniform spherical wave (NUSW) model, where UAV swarm trajectory simultaneously influences the channel amplitude and phase. We formulate a general optimization problem to maximize the minimum average communication rate over user equipments (UEs), by jointly optimizing the 3D UAV swarm trajectory and receive beamforming for all UEs. To draw useful insights, the special case of single UE is first studied, and successive convex approximation (SCA) technique is proposed to efficiently optimize the UAV swarm trajectory. For the special case of placement optimization, the optimal placement positions of UAVs for cases of single UAV and two UAVs are derived in closed-form. Then, for the special case of two UEs, we show that an inter-UE interference (IUI)-free communication can be achieved by symmetrically placing an even number of UAVs along a hyperbola, with its foci corresponding to the locations of the two UEs. Furthermore, for arbitrary number of UEs, an alternating optimization algorithm is proposed to efficiently tackle the non-convex optimization problem. Numerical results validate the significant performance gains over the benchmark schemes.

cs.IT↗

arXiv2026

Demonstration of High-Gain Harmonic Lasing in a Terahertz Free-Electron Laser

Compact Free-Electron Lasers (FELs) offering broad, continuous spectral tunability are traditionally constrained by fixed-parameter magnetic structures and the necessity for high-energy electron beams. High-gain Harmonic Lasing (HL) has long been proposed as a solution to overcome these limitations; however, a robust experimental verification of this principle has remained absent. Here, we report the first experimental demonstration of high-gain HL. By employing a frequency-tunable electron beam density modulation to dominate the fundamental instability, we achieved sustained FEL amplification at the 3rd and 5th harmonics of the wiggler. The HL mode generated output power comparable to conventional fundamental operation with enhanced stability and narrower spectral bandwidth. Notably, we demonstrate that HL extends the spectral coverage by a factor of two under fixed facility constraints, achieving pulse energies up to 540 μJ. These results establish high-gain HL as a versatile mechanism for advancing compact, wavelength-flexible FEL facilities.

physics.acc-ph↗

arXiv2026

SAIL-RL: Guiding MLLMs in When and How to Think via Dual-Reward RL Tuning

We introduce SAIL-RL, a reinforcement learning (RL) post-training framework that enhances the reasoning capabilities of multimodal large language models (MLLMs) by teaching them when and how to think. Existing approaches are limited by outcome-only supervision, which rewards correct answers without ensuring sound reasoning, and by uniform thinking strategies, which often lead to overthinking on simple tasks and underthinking on complex ones. SAIL-RL addresses these challenges with a dual reward system: the Thinking Reward, which evaluates reasoning quality through factual grounding, logical coherence, and answer consistency, and the Judging Reward, which adaptively determines whether deep reasoning or direct answering is appropriate. Experiments on the state-of-the-art SAIL-VL2 show that SAIL-RL improves reasoning and multimodal understanding benchmarks at both 4B and 8B scales, achieving competitive performance against commercial closed-source models such as GPT-4o, and substantially reduces hallucinations, establishing it as a principled framework for building more reliable and adaptive MLLMs. The code will be available at https://github.com/BytedanceDouyinContent/SAIL-RL.

cs.CV↗

arXiv2026

Learning Deep Tree-based Retriever for Efficient Recommendation: Theory and Method

Although advancements in deep learning have significantly enhanced the recommendation accuracy of deep recommendation models, these methods still suffer from low recommendation efficiency. Recently proposed tree-based deep recommendation models alleviate the problem by directly learning tree structure and representations under the guidance of recommendation objectives. To guarantee the effectiveness of beam search for recommendation accuracy, these models strive to ensure that the tree adheres to the max-heap assumption, where a parent node's preference should be the maximum among its children's preferences. However, they employ a one-versus-all strategy, framing the training task as a series of independent binary classification objectives for each node, which limits their ability to fully satisfy the max-heap assumption. To this end, we propose a Deep Tree-based Retriever (DTR for short) for efficient recommendation. DTR frames the training task as a softmax-based multi-class classification over tree nodes at the same level, enabling explicit horizontal competition and more discriminative top-k selection among them, which mimics the beam search behavior during training. To mitigate the suboptimality induced by the labeling of non-leaf nodes, we propose a rectification method for the loss function, which further aligns with the max-heap assumption in expectation. As the number of tree nodes grows exponentially with the levels, we employ sampled softmax to approximate optimization and thereby enhance efficiency. Furthermore, we propose a tree-based sampling method to reduce the bias inherent in sampled softmax. Theoretical results reveal DTR's generalization capability, and both the rectification method and tree-based sampling contribute to improved generalization. The experiments are conducted on four real-world datasets, validating the effectiveness of the proposed method.

cs.IR↗

arXiv2026

M$^3$Searcher: Modular Multimodal Information Seeking Agency with Retrieval-Oriented Reasoning

Recent advances in DeepResearch-style agents have demonstrated strong capabilities in autonomous information acquisition and synthesize from real-world web environments. However, existing approaches remain fundamentally limited to text modality. Extending autonomous information-seeking agents to multimodal settings introduces critical challenges: the specialization-generalization trade-off that emerges when training models for multimodal tool-use at scale, and the severe scarcity of training data capturing complex, multi-step multimodal search trajectories. To address these challenges, we propose M$^3$Searcher, a modular multimodal information-seeking agent that explicitly decouples information acquisition from answer derivation. M$^3$Searcher is optimized with a retrieval-oriented multi-objective reward that jointly encourages factual accuracy, reasoning soundness, and retrieval fidelity. In addition, we develop MMSearchVQA, a multimodal multi-hop dataset to support retrieval centric RL training. Experimental results demonstrate that M$^3$Searcher outperforms existing approaches, exhibits strong transfer adaptability and effective reasoning in complex multimodal tasks.

cs.AI↗

arXiv2025

AdaLRS: Loss-Guided Adaptive Learning Rate Search for Efficient Foundation Model Pretraining

Learning rate is widely regarded as crucial for effective foundation model pretraining. Recent research explores and demonstrates the transferability of learning rate configurations across varying model and dataset sizes, etc. Nevertheless, these approaches are constrained to specific training scenarios and typically necessitate extensive hyperparameter tuning on proxy models. In this work, we propose \textbf{AdaLRS}, a plug-in-and-play adaptive learning rate search algorithm that conducts online optimal learning rate search via optimizing loss descent velocities. We provide theoretical and experimental analyzes to show that foundation model pretraining loss and its descent velocity are both convex and share the same optimal learning rate. Relying solely on training loss dynamics, AdaLRS involves few extra computations to guide the search process, and its convergence is guaranteed via theoretical analysis. Experiments on both LLM and VLM pretraining show that AdaLRS adjusts suboptimal learning rates to the neighborhood of optimum with marked efficiency and effectiveness, with model performance improved accordingly. We also show the robust generalizability of AdaLRS across varying training scenarios, such as different model sizes, training paradigms, base learning rate scheduler choices, and hyperparameter settings.

cs.LG↗

arXiv2025

MEML-GRPO: Heterogeneous Multi-Expert Mutual Learning for RLVR Advancement

Recent advances demonstrate that reinforcement learning with verifiable rewards (RLVR) significantly enhances the reasoning capabilities of large language models (LLMs). However, standard RLVR faces challenges with reward sparsity, where zero rewards from consistently incorrect candidate answers provide no learning signal, particularly in challenging tasks. To address this, we propose Multi-Expert Mutual Learning GRPO (MEML-GRPO), an innovative framework that utilizes diverse expert prompts as system prompts to generate a broader range of responses, substantially increasing the likelihood of identifying correct solutions. Additionally, we introduce an inter-expert mutual learning mechanism that facilitates knowledge sharing and transfer among experts, further boosting the model's performance through RLVR. Extensive experiments across multiple reasoning benchmarks show that MEML-GRPO delivers significant improvements, achieving an average performance gain of 4.89% with Qwen and 11.33% with Llama, effectively overcoming the core limitations of traditional RLVR methods.

cs.AI↗

arXiv2025

Semi-Supervised Synthetic Data Generation with Fine-Grained Relevance Control for Short Video Search Relevance Modeling

Synthetic data is widely adopted in embedding models to ensure diversity in training data distributions across dimensions such as difficulty, length, and language. However, existing prompt-based synthesis methods struggle to capture domain-specific data distributions, particularly in data-scarce domains, and often overlook fine-grained relevance diversity. In this paper, we present a Chinese short video dataset with 4-level relevance annotations, filling a critical resource void. Further, we propose a semi-supervised synthetic data pipeline where two collaboratively trained models generate domain-adaptive short video data with controllable relevance labels. Our method enhances relevance-level diversity by synthesizing samples for underrepresented intermediate relevance labels, resulting in a more balanced and semantically rich training data set. Extensive offline experiments show that the embedding model trained on our synthesized data outperforms those using data generated based on prompting or vanilla supervised fine-tuning(SFT). Moreover, we demonstrate that incorporating more diverse fine-grained relevance levels in training data enhances the model's sensitivity to subtle semantic distinctions, highlighting the value of fine-grained relevance supervision in embedding learning. In the search enhanced recommendation pipeline of Douyin's dual-column scenario, through online A/B testing, the proposed model increased click-through rate(CTR) by 1.45%, raised the proportion of Strong Relevance Ratio (SRR) by 4.9%, and improved the Image User Penetration Rate (IUPR) by 0.1054%.

cs.CL↗

arXiv2025

Smarter, not Bigger: Fine-Tuned RAG-Enhanced LLMs for Automotive HIL Testing

Hardware-in-the-Loop (HIL) testing is essential for automotive validation but suffers from fragmented and underutilized test artifacts. This paper presents HIL-GPT, a retrieval-augmented generation (RAG) system integrating domain-adapted large language models (LLMs) with semantic retrieval. HIL-GPT leverages embedding fine-tuning using a domain-specific dataset constructed via heuristic mining and LLM-assisted synthesis, combined with vector indexing for scalable, traceable test case and requirement retrieval. Experiments show that fine-tuned compact models, such as \texttt{bge-base-en-v1.5}, achieve a superior trade-off between accuracy, latency, and cost compared to larger models, challenging the notion that bigger is always better. An A/B user study further confirms that RAG-enhanced assistants improve perceived helpfulness, truthfulness, and satisfaction over general-purpose LLMs. These findings provide insights for deploying efficient, domain-aligned LLM-based assistants in industrial HIL environments.

cs.CL↗

arXiv2025

Lattice design of a storage-ring-based light source for generating high-power fully coherent EUV radiation

We present the physical design and systematic optimization of a high-performance storage ring tailored for the generation of high-power coherent radiation, with particular emphasis on the extreme ultraviolet (EUV) regime. The proposed ring adopts a Double Bend Achromat (DBA) lattice configuration and integrates 12 superconducting wigglers to significantly enhance radiation damping and minimize the natural emittance. And a bypass line is adopted to generate high power coherent radiation. Comprehensive linear and nonlinear beam dynamics analyses have been conducted to ensure beam stability and robustness across the operational parameter space. The optimized design achieves a natural emittance of approximately 0.8 nm and a longitudinal damping time of around 1.4 ms, enabling the efficient buildup of coherent radiation. Three-dimensional numerical simulations, incorporating the previously proposed angular dispersion-induced microbunching (ADM) mechanism, further confirm the system's capability to generate high-power EUV coherent radiation, with output powers reaching the order of several hundred watts. These results underscore the strong potential of the proposed design for applications in coherent photon science and EUV lithography.

physics.acc-ph↗

arXiv2025

Understanding and Optimizing Agentic Workflows via Shapley value

Agentic workflows have become the dominant paradigm for building complex AI systems, orchestrating specialized components, such as planning, reasoning, action execution, and reflection, to tackle sophisticated real-world tasks. However, systematically analyzing and optimizing these workflows remains challenging due to intricate component interdependencies and the lack of principled attribution methods. In this work, we introduce ShapleyFlow, the first framework that employs cooperative game theory to analyze and optimize agentic workflows. By applying the Shapley value to evaluate all possible component configurations, ShapleyFlow enables fine-grained attribution of each component's contribution and facilitates the identification of task-specific optimal configurations. Through a constructed dataset evaluated across 7 scenarios, such as navigation, math and OS, we demonstrate 3 key contributions: (1) Theoretical Framework: a principled game-theoretic approach for the attribution of contributions in agentic workflows. (2) Optimal Workflow Discovery: ShapleyFlow identifies task-specific component configurations that consistently outperform workflows relying on a single LLM across all tested tasks. (3) Comprehensive Analysis: we construct and analyze over 1,500 tasks, providing actionable insights and design guidelines for optimizing workflows across multiple domains.

cs.AI↗