Searcharxiv⌕ Search

arXiv subjects

Heming Cui

Publications and source records attributed to Heming Cui.

At least 19 recordsLinked to original sources
arXiv2026

FluxShard: Motion-Aware Feature Cache Reuse for Collaborative Video Analytics in Mobile Edge Computing

Caching and reusing intermediate features across consecutive frames is a common technique to reduce redundant computation and transmission for edge-cloud video analytics in mobile edge computation. Existing methods manage the cache in a fixed or globally shifted coordinate system, treating it as an indivisible whole. Under the non-uniform motion patterns of mobile scenes, this whole-scene granularity invalidates large portions of the cache even when most content has merely shifted spatially, wasting computation and bandwidth. The root cause is a granularity mismatch: the cache is managed per scene, yet motion varies per region. In this paper, we present FluxShard, a motion-aware edge-cloud video analytics system that uses codec-level block motion vectors (MVs) to manage feature cache reuse and recomputation at the granularity of individual motion regions. By re-indexing cached features along per-block MVs, FluxShard separates spatial displacement from content changes, recovering reusable content that whole-scene methods would otherwise discard. To ensure correct reuse under heterogeneous motion, the Receptive Field Alignment Principle (RFAP) identifies, from the input-level MV field alone, the positions that must be recomputed due to inconsistent spatial composition within receptive fields. To maintain cache coherence across frames, MV-guided cache remapping warps the entire feature cache to the current coordinate system each frame, sustaining a high reuse ratio over time. A profiling-driven dispatcher routes the remaining sparse workload between edge and cloud for lower latency. Evaluation across multiple vision tasks, dynamic video benchmarks, and network conditions shows that FluxShard reduces latency by 32.6-83.8% and energy by 14.9-64.0% over all baselines under the prescribed accuracy budget.

cs.NI↗
arXiv2026

Transformer-Based Multipath Congestion Control: A Decoupled Approach for Wireless Uplinks

The proliferation of artificial intelligence applications on edge devices necessitates efficient transport protocols that leverage multi-homed connectivity across heterogeneous networks. While Multipath TCP enables bandwidth aggregation, its in-kernel congestion control mechanisms lack the programmability and flexibility needed for achieving efficient transmission. Additionally, inherent measurement noise renders network state partially observable, challenging data-driven approaches like deep reinforcement learning (DRL). To address these challenges, we propose a Transformer-based Congestion Control Optimization (TCCO) framework for multipath transport. TCCO employs a decoupled architecture that offloads control decisions to an external decision engine via a lightweight in-kernel client and user-space proxy, enabling edge devices to leverage external computational resources while maintaining TCP/IP compatibility. The Transformer-based DRL agent in the external decision engine uses self-attention to capture temporal dependencies, filter noise, and coordinate control across subflows through a unified policy. Extensive evaluation on both simulated and real dual-band Wi-Fi testbeds demonstrates that TCCO achieves superior adaptability and performance than state-of-the-art baselines, validating the feasibility and effectiveness of TCCO for wireless networks.

cs.NI↗
arXiv2025

StereoVLA: Enhancing Vision-Language-Action Models with Stereo Vision

While Vision-Language-Action (VLA) models excel in generalist manipulation, they often lack fine-grained spatial awareness and show limited viewpoint robustness. This limitation largely stems from the reliance on pretrained RGB encoders, which lack explicit geometric cues and prioritize semantic alignment over geometric representation. We argue that effective visual representations for VLA models must jointly encode both semantic and geometric information. In this paper, we introduce StereoVLA, the first VLA model to incorporate rich geometric cues from large-scale synthetic stereo data. StereoVLA employs a Geometric-and-Semantic (GeoSem) vision encoder that extracts geometric cues from subtle stereo-view disparities for precise spatial perception, while simultaneously capturing semantic features from pixel observations to support language-conditioned manipulation. Additionally, we introduce two synergistic co-training objectives: Interaction-Region Depth Estimation for precise spatial reasoning, and Camera Parameter Estimation to implicitly align perception and action coordinate systems. Compared with baselines that employ various input modalities, StereoVLA achieves a 33.4% absolute gain in success rate in real-world experiments and demonstrates robustness to near-hemispheric camera perspectives. Project page: https://shengliangd.github.io/StereoVLA-Webpage.

cs.RO↗
arXiv2025

LLM-Driven Stationarity-Aware Expert Demonstrations for Multi-Agent Reinforcement Learning in Mobile Systems

Multi-agent reinforcement learning (MARL) has been increasingly adopted in many real-world applications. While MARL enables decentralized deployment on resource-constrained edge devices, it suffers from severe non-stationarity due to the synchronous updates of agent policies. This non stationarity results in unstable training and poor policy con vergence, especially as the number of agents increases. In this paper, we propose RELED, a scalable MARL framework that integrates large language model (LLM)-driven expert demonstrations with autonomous agent exploration. RELED incorporates a Stationarity-Aware Expert Demonstration module, which leverages theoretical non-stationarity bounds to enhance the quality of LLM-generated expert trajectories, thus providing high reward and training-stable samples for each agent. Moreover, a Hybrid Expert-Agent Policy Optimization module adaptively balances each agent's learning from both expert-generated and agent-generated trajectories, accelerating policy convergence and improving generalization. Extensive experiments with real city networks based on OpenStreetMap demonstrate that RELED achieves superior performance compared to state-of-the-art MARL methods.

cs.LG↗
arXiv2025

LEED: A Highly Efficient and Scalable LLM-Empowered Expert Demonstrations Framework for Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning (MARL) holds substantial promise for intelligent decision-making in complex environments. However, it suffers from a coordination and scalability bottleneck as the number of agents increases. To address these issues, we propose the LLM-empowered expert demonstrations framework for multi-agent reinforcement learning (LEED). LEED consists of two components: a demonstration generation (DG) module and a policy optimization (PO) module. Specifically, the DG module leverages large language models to generate instructions for interacting with the environment, thereby producing high-quality demonstrations. The PO module adopts a decentralized training paradigm, where each agent utilizes the generated demonstrations to construct an expert policy loss, which is then integrated with its own policy loss. This enables each agent to effectively personalize and optimize its local policy based on both expert knowledge and individual experience. Experimental results show that LEED achieves superior sample efficiency, time efficiency, and robust scalability compared to state-of-the-art baselines.

cs.MA↗
arXiv2025

Sample Efficient Experience Replay in Non-stationary Environments

Reinforcement learning (RL) in non-stationary environments is challenging, as changing dynamics and rewards quickly make past experiences outdated. Traditional experience replay (ER) methods, especially those using TD-error prioritization, struggle to distinguish between changes caused by the agent's policy and those from the environment, resulting in inefficient learning under dynamic conditions. To address this challenge, we propose the Discrepancy of Environment Dynamics (DoE), a metric that isolates the effects of environment shifts on value functions. Building on this, we introduce Discrepancy of Environment Prioritized Experience Replay (DEER), an adaptive ER framework that prioritizes transitions based on both policy updates and environmental changes. DEER uses a binary classifier to detect environment changes and applies distinct prioritization strategies before and after each shift, enabling more sample-efficient learning. Experiments on four non-stationary benchmarks demonstrate that DEER further improves the performance of off-policy algorithms by 11.54 percent compared to the best-performing state-of-the-art ER methods.

cs.LG↗
arXiv2025

RRTO: A High-Performance Transparent Offloading System for Model Inference in Mobile Edge Computing

Deploying Machine Learning (ML) applications on resource-constrained mobile devices remains challenging due to limited computational resources and poor platform compatibility. While Mobile Edge Computing (MEC) offers offloading-based inference paradigm using GPU servers, existing approaches are divided into non-transparent and transparent methods, with the latter necessitating modifications to the source code. Non-transparent offloading achieves high performance but requires intrusive code modification, limiting compatibility with diverse applications. Transparent offloading, in contrast, offers wide compatibility but introduces significant transmission delays due to per-operator remote procedure calls (RPCs). To overcome this limitation, we propose RRTO, the first high-performance transparent offloading system tailored for MEC inference. RRTO introduces a record/replay mechanism that leverages the static operator sequence in ML models to eliminate repetitive RPCs. To reliably identify this sequence, RRTO integrates a novel Operator Sequence Search algorithm that detects repeated patterns, filters initialization noise, and accelerates matching via a two-level strategy. Evaluation demonstrates that RRTO achieves substantial reductions of up to 98% in both per-inference latency and energy consumption compared to state-of-the-art transparent methods and yields results comparable to non-transparent approaches, all without necessitating any source code modification.

cs.NI↗
arXiv2025

Intra-DP: A High Performance Collaborative Inference System for Mobile Edge Computing

Deploying deep neural networks (DNNs) on resource-constrained mobile devices presents significant challenges, particularly in achieving real-time performance while simultaneously coping with limited computational resources and battery life. While Mobile Edge Computing (MEC) offers collaborative inference with GPU servers as a promising solution, existing approaches primarily rely on layer-wise model partitioning and undergo significant transmission bottlenecks caused by the sequential execution of DNN operations. To address this challenge, we present Intra-DP, a high-performance collaborative inference system optimized for DNN inference on MEC. Intra DP employs a novel parallel computing technique based on local operators (i.e., operators whose minimum unit input is not the entire input tensor, such as the convolution kernel). By decomposing their computations (operations) into several independent sub-operations and overlapping the computation and transmission of different sub-operations through parallel execution, Intra-DP mitigates transmission bottlenecks in MEC, achieving fast and energy-efficient inference. The evaluation demonstrates that Intra-DP reduces per-inference latency by up to 50% and energy consumption by up to 75% compared to state-of-the-art baselines, without sacrificing accuracy.

cs.NI↗
arXiv2025

EffiBench-X: A Multi-Language Benchmark for Measuring Efficiency of LLM-Generated Code

Existing code generation benchmarks primarily evaluate functional correctness, with limited focus on code efficiency and often restricted to a single language like Python. To address this gap, we introduce EffiBench-X, the first multi-language benchmark designed to measure the efficiency of LLM-generated code. EffiBench-X supports Python, C++, Java, JavaScript, Ruby, and Golang. It comprises competitive programming tasks with human-expert solutions as efficiency baselines. Evaluating state-of-the-art LLMs on EffiBench-X reveals that while models generate functionally correct code, they consistently underperform human experts in efficiency. Even the most efficient LLM-generated solutions (Qwen3-32B) achieve only around \textbf{62\%} of human efficiency on average, with significant language-specific variations. LLMs show better efficiency in Python, Ruby, and JavaScript than in Java, C++, and Golang. For instance, DeepSeek-R1's Python code is significantly more efficient than its Java code. These results highlight the critical need for research into LLM optimization techniques to improve code efficiency across diverse languages. The dataset and evaluation infrastructure are submitted and available at https://github.com/EffiBench/EffiBench-X.git and https://huggingface.co/datasets/EffiBench/effibench-x.

cs.CL↗
arXiv2025

GraspVLA: a Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data

Embodied foundation models are gaining increasing attention for their zero-shot generalization, scalability, and adaptability to new tasks through few-shot post-training. However, existing models rely heavily on real-world data, which is costly and labor-intensive to collect. Synthetic data offers a cost-effective alternative, yet its potential remains largely underexplored. To bridge this gap, we explore the feasibility of training Vision-Language-Action models entirely with large-scale synthetic action data. We curate SynGrasp-1B, a billion-frame robotic grasping dataset generated in simulation with photorealistic rendering and extensive domain randomization. Building on this, we present GraspVLA, a VLA model pretrained on large-scale synthetic action data as a foundational model for grasping tasks. GraspVLA integrates autoregressive perception tasks and flow-matching-based action generation into a unified Chain-of-Thought process, enabling joint training on synthetic action data and Internet semantics data. This design helps mitigate sim-to-real gaps and facilitates the transfer of learned actions to a broader range of Internet-covered objects, achieving open-vocabulary generalization in grasping. Extensive evaluations across real-world and simulation benchmarks demonstrate GraspVLA's advanced zero-shot generalizability and few-shot adaptability to specific human preferences. We will release SynGrasp-1B dataset and pre-trained weights to benefit the community.

cs.RO↗
arXiv2025

K2: On Optimizing Distributed Transactions in a Multi-region Data Store with TrueTime Clocks (Extended Version)

TrueTime clocks (TTCs) that offer accurate and reliable time within limited uncertainty bounds have been increasingly implemented in many clouds. Multi-region data stores that seek decentralized synchronization for high performance represent an ideal application of TTC. However, the co-designs between the two were often undervalued or failed to realize their full potential. This paper proposes K2, a multi-region data store that intensely explores the opportunity of using TTC for distributed transactions. Compared to its pioneer, Google Spanner, K2 augments TTC's semantics in three core design pillars. First, K2 carries a new timestamp-generating scheme that is capable of providing a small time uncertainty bound at scale. Second, K2 revitalizes existing multi-version timestamp-ordered concurrency control to realize multi-version properties for read-write transactions. Third, K2 introduces a new TTC-based visibility control protocol that provides efficient reads at replicas. Our evaluation shows that, K2 achieves an order of magnitude higher transaction throughput relative to other practical geo-distributed transaction protocols while ensuring a lower visibility delay at asynchronous replicas.

cs.DB↗
arXiv2025

State-Aware Perturbation Optimization for Robust Deep Reinforcement Learning

Recently, deep reinforcement learning (DRL) has emerged as a promising approach for robotic control. However, the deployment of DRL in real-world robots is hindered by its sensitivity to environmental perturbations. While existing whitebox adversarial attacks rely on local gradient information and apply uniform perturbations across all states to evaluate DRL robustness, they fail to account for temporal dynamics and state-specific vulnerabilities. To combat the above challenge, we first conduct a theoretical analysis of white-box attacks in DRL by establishing the adversarial victim-dynamics Markov decision process (AVD-MDP), to derive the necessary and sufficient conditions for a successful attack. Based on this, we propose a selective state-aware reinforcement adversarial attack method, named STAR, to optimize perturbation stealthiness and state visitation dispersion. STAR first employs a soft mask-based state-targeting mechanism to minimize redundant perturbations, enhancing stealthiness and attack effectiveness. Then, it incorporates an information-theoretic optimization objective to maximize mutual information between perturbations, environmental states, and victim actions, ensuring a dispersed state-visitation distribution that steers the victim agent into vulnerable states for maximum return reduction. Extensive experiments demonstrate that STAR outperforms state-of-the-art benchmarks.

cs.LG↗
arXiv2025

Robust Deep Reinforcement Learning in Robotics via Adaptive Gradient-Masked Adversarial Attacks

Deep reinforcement learning (DRL) has emerged as a promising approach for robotic control, but its realworld deployment remains challenging due to its vulnerability to environmental perturbations. Existing white-box adversarial attack methods, adapted from supervised learning, fail to effectively target DRL agents as they overlook temporal dynamics and indiscriminately perturb all state dimensions, limiting their impact on long-term rewards. To address these challenges, we propose the Adaptive Gradient-Masked Reinforcement (AGMR) Attack, a white-box attack method that combines DRL with a gradient-based soft masking mechanism to dynamically identify critical state dimensions and optimize adversarial policies. AGMR selectively allocates perturbations to the most impactful state features and incorporates a dynamic adjustment mechanism to balance exploration and exploitation during training. Extensive experiments demonstrate that AGMR outperforms state-of-the-art adversarial attack methods in degrading the performance of the victim agent and enhances the victim agent's robustness through adversarial defense mechanisms.

cs.LG↗
arXiv2025

Themis: Efficient Sparse Model Training Through Fully Sharded Sparse Data Parallelism

Mixture-of-Experts (MoE) scales large language models cost-effectively, but expert-parallel training suffers severe straggler effects from skewed expert loads. Current systems frequently rearrange expert placement to mitigate stragglers, inflating memory footprint and migration overhead, potentially negating the benefits of load balancing. We present Fully Sharded Sparse Data Parallelism (FSSDP), a sparse-native MoE training approach that enables in-situ load balancing on every training iteration, overlapping the balancing with computation and eliminating explicit expert rearrangement together with its migration traffic and memory reserves. FSSDP keeps MoE layers sharded and sparsely materializes an ephemeral, load-balancing placement each iteration, with re-materialization to reuse available memory across layers. FSSDP is complemented by heterogeneous sharding to shift memory imbalance from the device level to the layer level, maintaining uniform memory budgets while enabling per-layer placement optimization. We realize FSSDP in Themis with co-designed topology-aware placement algorithms. Across 2 clusters and diverse workloads, Themis achieves 1.26-2.42x speedup over state-of-the-art expert-rearrangement systems.

cs.DC↗
arXiv2025

Rethinking Adversarial Attacks in Reinforcement Learning from Policy Distribution Perspective

Deep Reinforcement Learning (DRL) suffers from uncertainties and inaccuracies in the observation signal in realworld applications. Adversarial attack is an effective method for evaluating the robustness of DRL agents. However, existing attack methods targeting individual sampled actions have limited impacts on the overall policy distribution, particularly in continuous action spaces. To address these limitations, we propose the Distribution-Aware Projected Gradient Descent attack (DAPGD). DAPGD uses distribution similarity as the gradient perturbation input to attack the policy network, which leverages the entire policy distribution rather than relying on individual samples. We utilize the Bhattacharyya distance in DAPGD to measure policy similarity, enabling sensitive detection of subtle but critical differences between probability distributions. Our experiment results demonstrate that DAPGD achieves SOTA results compared to the baselines in three robot navigation tasks, achieving an average 22.03% higher reward drop compared to the best baseline.

cs.LG↗
arXiv2024

EffiCoder: Enhancing Code Generation in Large Language Models through Efficiency-Aware Fine-tuning

As large language models (LLMs) play an increasingly important role in code generation, enhancing both correctness and efficiency has become crucial. Current methods primarily focus on correctness, often overlooking efficiency. To address this gap, we introduce EffiCoder to improve both aspects by fine-tuning LLMs on a high-quality dataset comprising correct and efficient code samples. Our methodology involves leveraging multiple LLMs to generate diverse candidate code solutions for various tasks across different programming languages. We then evaluate these solutions by measuring their execution time and memory usage through local execution. The code solution with the lowest execution time and memory consumption is selected as the final output for each task. Experimental results demonstrate significant improvements when fine-tuning with Effi-Instruct. For instance, Qwen2.5-Coder-7B-Instruct's pass@1 score increases from 44.8\% to 57.7\%, while the average execution time for correct tasks decreases by 48.4\%. EffiCoder offers a scalable and effective solution for advancing AI-driven code generation, benefiting software development and computational problem-solving. The source code of Effi-Code was released at https://github.com/huangd1999/EffiCoder.

cs.CL↗
arXiv2024

Towards Synergistic, Generalized, and Efficient Dual-System for Robotic Manipulation

The increasing demand for versatile robotic systems to operate in diverse and dynamic environments has emphasized the importance of a generalist policy, which leverages a large cross-embodiment data corpus to facilitate broad adaptability and high-level reasoning. However, the generalist would struggle with inefficient inference and cost-expensive training. The specialist policy, instead, is curated for specific domain data and excels at task-level precision with efficiency. Yet, it lacks the generalization capacity for a wide range of applications. Inspired by these observations, we introduce RoboDual, a synergistic dual-system that supplements the merits of both generalist and specialist policy. A diffusion transformer-based specialist is devised for multi-step action rollouts, exquisitely conditioned on the high-level task understanding and discretized action output of a vision-language-action (VLA) based generalist. Compared to OpenVLA, RoboDual achieves 26.7% improvement in real-world setting and 12% gain on CALVIN by introducing a specialist policy with merely 20M trainable parameters. It maintains strong performance with 5% of demonstration data only, and enables a 3.8 times higher control frequency in real-world deployment. Code would be made publicly available. Our project page is hosted at: https://opendrivelab.com/RoboDual/

cs.RO↗
arXiv2024

HE-Nav: A High-Performance and Efficient Navigation System for Aerial-Ground Robots in Cluttered Environments

Existing AGR navigation systems have advanced in lightly occluded scenarios (e.g., buildings) by employing 3D semantic scene completion networks for voxel occupancy prediction and constructing Euclidean Signed Distance Field (ESDF) maps for collision-free path planning. However, these systems exhibit suboptimal performance and efficiency in cluttered environments with severe occlusions (e.g., dense forests or tall walls), due to limitations arising from perception networks' low prediction accuracy and path planners' high computational overhead. In this paper, we present HE-Nav, the first high-performance and efficient navigation system tailored for AGRs operating in cluttered environments. The perception module utilizes a lightweight semantic scene completion network (LBSCNet), guided by a bird's eye view (BEV) feature fusion and enhanced by an exquisitely designed SCB-Fusion module and attention mechanism. This enables real-time and efficient obstacle prediction in cluttered areas, generating a complete local map. Building upon this completed map, our novel AG-Planner employs the energy-efficient kinodynamic A* search algorithm to guarantee planning is energy-saving. Subsequent trajectory optimization processes yield safe, smooth, dynamically feasible and ESDF-free aerial-ground hybrid paths. Extensive experiments demonstrate that HE-Nav achieved 7x energy savings in real-world situations while maintaining planning success rates of 98% in simulation scenarios. Code and video are available on our project page: https://jmwang0117.github.io/HE-Nav/.

cs.RO↗