arXiv subjects

Jun Ma

Publications and source records attributed to Jun Ma.

At least 37 records · Page 2Linked to original sources

arXiv2026

EmoPose: Vision-Language Model Guided Emotion-Aware Gesture Generation for Humanoid Robots

Socially competent humanoid robots must communicate affect and intent through gesture as well as speech, yet open-ended interaction must become motion that is both expressive and executable on a specific body. This demands semantic flexibility for contextual social intent while preserving deterministic, embodiment-aware robot control. We present EmoPose, a vision-language model (VLM)-guided framework that bridges this gap through an executable semantic interface. Given language, dialogue history, and optional visual context, the VLM selects an ordered gesture plan containing a communicative class, library variant, intensity, and speech anchor. A scalable robot-owned motion library defines the available expressive vocabulary and the source of 14-DoF joint targets. Pose Studio supports automatic trajectory generation, MuJoCo preview, and automatic synchronization of new library entries with the VLM guide; deterministic robot-side modules validate plans, construct trajectories, schedule gestures, and manage queueing and interruption. This division lets the interaction repertoire grow for new social contexts without changing the control interface or delegating raw joint commands to the foundation model. On the EmoPose-Bench, structured GPT-5.5 planning reaches $98.25\pm0.52\%$ on the Easy tier and $76.50\pm0.54\%$ overall, exceeding same-model direct-label prompting. Further tests validate dialogue-context use and ordered multi-action composition. The system completes the nominal MuJoCo suite and realizes all 29 authored variants on the physical Unitree G1. A four-stop laboratory tour demonstrates expressive narration with interruption, camera-grounded dialogue, and navigation.

cs.RO↗

arXiv2026

ForceRFT: Refining VLA Actions through Force-Guided Residual Reinforcement Learning

Force-conditioned vision-language-action (VLA) policies can respond to contact, but when trained solely on demonstrations, their recovery behavior may be limited by demonstration coverage, and they do not learn from deployment outcomes. Human corrective imitation provides additional recovery examples, but its objective matches local action targets without explicitly optimizing task return. We present ForceRFT, a force-guided residual reinforcement learning framework that learns contact-dependent corrections from human supervision and autonomous task outcomes. A frozen, demonstration-trained SmolVLA-based prior generates force-conditioned action chunks, while a lightweight residual actor refines individual end-effector pose commands using wrist feedback acquired during chunk execution. The decision-time wrist wrench, its temporal change, and the selected base motion condition both residual correction and value estimation. Human corrections supervise the residual actor, while verified autonomous transitions train the twin critics and support value-guided updates to the same actor. Bootstrapping is restricted to autonomous segments, preventing TD credit from crossing human-intervention boundaries. Real-robot experiments on plug insertion, ring-on-peg assembly, and whiteboard wiping show higher autonomous success rates than the evaluated demonstration-trained and residual-imitation baselines. Comparisons with residual imitation support value-guided residual optimization, while plug-insertion ablations indicate the benefit of direct execution-time wrist feedback.

cs.RO↗

arXiv2026

SFVO: Decoupled Confidence-Guided Stereo-Flow Visual Odometry with Bidirectional PnP

Deep learning-based visual odometry (VO) has achieved significant progress, yet most existing methods focus on a monocular approach, which suffers from scale ambiguity. Stereo VO provides real metric by its nature, but remains less studied in deep learning VO due to its high computational cost and modeling complexity. Recent advances in stereo matching and optical flow estimation have made dense visual correspondence increasingly accurate and reliable, but their complementary geometric information has not been fully exploited for VO. In this paper, we present SFVO, a correspondence-driven stereo VO framework that directly builds upon pretrained stereo matching and optical flow models. SFVO exploits pretrained stereo matching and optical flow models to estimate stereo and temporal correspondences. Instead of learning pose directly from images, SFVO maps learned correspondences into geometric constraints and predicts which points are trustworthy. To improve the reliability of visual correspondence-based geometric constraints, we introduce decoupled confidence maps for rotation and translation. This design better aligns the characteristics of visual correspondence and 6-DoF transformations. Extensive experiments on outdoor and indoor datasets demonstrate that SFVO achieves robust and accurate pose estimation with strong generalization capability. The code will be released.

cs.CV↗

arXiv2026

GR2PO: Group Relative Return Policy Optimization for Continuous Robot Control

Actor-critic architecture has been widely used in continuous robot control. However, they rely on learning a value network, introducing additional computational overhead during training. Moreover, policy learning may also be affected by the approximation error of value estimation. Critic-free group relative policy optimization methods provide a simpler training approach by removing the need for a critic. However, they fail to learn long-term action outcomes when directly applying immediate rewards to policy optimization in dense-reward environments. To address these problems, we propose Group Relative Return Policy Optimization (GR2PO), a critic-free reinforcement learning framework for continuous robot control. GR2PO estimates the discounted returns from the parallelly collected trajectories, performs group normalization at each rollout time index, and uses relative advantages and clipped targets to update the policy. To evaluate the effectiveness of the proposed framework, we instantiate it on robot control simulation environments and deploy the model to a real-world edge device. The results show that GR2PO significantly outperforms critic-free baselines that use immediate rewards and performs competitively against state-of-the-art actor-critic methods. Furthermore, GR2PO demonstrates competitive training efficiency. Inference tests on NVIDIA Jetson TX2 demonstrate the feasibility of deploying the learned policies on edge platforms. Further ablation experiments analyze the effects of parallel group size, return estimation methods, and target clipping ratio on learning performance. To support follow-up research, we will make the complete code publicly available after the paper is accepted, including the framework implementation, experimental configuration, and training and evaluation scripts.

cs.RO↗

arXiv2026

Learning Reliable Parking Policies via Offline Reinforcement Learning with Quantized Action Representations

Parking is a routine yet safety-critical task for autonomous vehicles operating in urban environments. However, cluttered and weakly structured parking spaces, compounded by the interactive uncertainty from surrounding vehicles, hinder reliable maneuver generation. To address these challenges, we develop a waypoint-level offline reinforcement learning framework for interaction-aware autonomous parking. Specifically, a dedicated parking dataset is constructed from hierarchical expert rollouts with rotational waypoint augmentation, covering both non-interactive scenarios and interactive ones. The policy is then conditioned on a compact state representation, in which LiDAR-based obstacle features are adapted to the target pose via feature-wise linear modulation. A state-conditioned tokenizer further quantizes continuous waypoint sequences into discrete action tokens, over which conservative Q-learning is performed to suppress value overestimation on poorly supported actions. Extensive closed-loop experiments are conducted in the high-fidelity CARLA simulator. The proposed framework attains the highest parking success rate among all baselines and transfers reliably to unseen parking slots.

cs.RO↗

arXiv2026

LiteViLNet: Lightweight Vision-LiDAR Fusion Network for Efficient Road Segmentation

Road segmentation is a fundamental perception task for autonomous driving and mobile robotics, where both appearance and geometric cues must be processed under edge-computing constraints. Existing multi-modal approaches often improve accuracy with large encoders or expensive global interaction, which limits their use on embedded platforms. We present \textbf{LiteViLNet}, a lightweight RGB-geometry fusion network that combines a MobileNetV3 RGB encoder with a 0.12M-parameter depth-wise-separable geometry encoder. A multi-scale feature fusion module performs modality-specific enhancement, global-query cross-modal interaction, and adaptive gating, while a depth-wise large-kernel bridge enlarges the contextual support of the deepest representation with low overhead. The resulting U-Net-style decoder uses deep supervision only during training. On the KITTI Road benchmark, the 14.04M-parameter full model obtains $97.23\pm0.15\%$ MaxF. On the held-out ORFD test set under the released OFF-Net evaluation protocol, the full model achieves $96.74\pm0.09\%$ F-score and $93.68\pm0.18\%$ IoU. On a Jetson Orin NX, model-only PyTorch FP16 inference reaches $22.18\pm0.21$ FPS; a separate TensorRT FP16 measurement reaches $68.73\pm0.06$ FPS on the Jetson. Camera-depth adaptations and perception-and-control demonstrations on three heterogeneous robot platforms further illustrate the portability of the dual-stream design.

cs.CV↗

arXiv2026

Latency-Tolerant Cloud-Edge Collaborative Vision-Language-Action Models via Emergent Representational Specialization

Deploying billion-parameter Vision-Language-Action (VLA) policies on mobile robots creates a systems conflict: semantic reasoning benefits from cloud GPUs, whereas closed-loop control must respond locally despite network delay and jitter. Existing hierarchical and asynchronous policies improve throughput, but their slow-path representations can still arrive stale or require explicit scheduling and delay cues. We introduce CloudEdgeVLA, a cloud-edge policy that treats temporal misalignment as a representation-learning problem. A cloud VLA encodes delayed observations into slowly varying task features, while a lightweight edge head combines the latest available cloud feature with current local vision. During training, current and randomly delayed frames are paired with the same current action target in fresh and stale paths. This objective encourages the cloud representation to preserve task-level information while the edge path supplies state-sensitive corrections, driving emergent specialization. Across four LIBERO suites, CloudEdgeVLA retains 63.8-78.0% success with a 40-step uniform-delay window, whereas VLASH reaches at most 6.4% and the evaluated single-path baselines at most 3.0%. By removing blocking synchronization from the control loop, the design offers a practical route to scalable VLA deployment in which cloud models can grow while edge computation remains lightweight and responsive.

cs.RO↗

arXiv2026

Function-Preserving Data Generation for Zero-Shot Real-to-Sim-to-Real Manipulation

Robotic data generation is a promising paradigm for scaling robot learning without collecting large-scale real-world data. However, generating geometrically diverse yet physically valid data for contact-rich tasks remains challenging, especially when success depends on precise geometric interfaces. Standard shape augmentation methods often distort task-critical interfaces, resulting in invalid contact relationships, e.g., fit mismatches or interpenetration, rendering downstream interactions infeasible. To address these limitations, we propose a function-preserving Real-to-Sim-to-Real framework that generates synthetic demonstrations from reconstructed assets without teleoperated source trajectories. Our method augments task-relevant object geometries through constraint-guided mesh deformation, together with physically consistent transfer of task poses and collision proxies. Visual domain randomization is further applied during simulation rollouts, enabling robust zero-shot policy deployment without real-world fine-tuning. Extensive experiments in both real-world and simulation settings demonstrate that our method enables robust generalization across unseen object geometries and diverse visual conditions in contact-rich and long-horizon tasks. Our method provides a practical path toward scalable robot learning for contact-rich tasks via shape deformation.

cs.RO↗

arXiv2026

A general lightweight global modeling framework for three-dimensional seismic exploration

In seismic exploration, the propagation of seismic waves naturally gives rise to long-range dependencies in seismic data. Capturing such global correlations can significantly improve the accuracies of seismic signal processing, inversion, and interpretation. Global modeling (GM) methods have therefore emerged as an effective paradigm for seismic exploration, offering a powerful means of exploiting the intrinsic global relationships within seismic data. However, the mainstream GM approaches, particularly Transformers, incur prohibitively high computational costs for their inherent quadratic complexity, which restricts the scalability of such methods to higher-dimensional data. To achieve efficient GM for three-dimensional (3D) seismic exploration, we propose a general framework, named lightweight Mamba-based global modeling (LMGM), to establish long-range dependencies at a relatively low computational cost. Specifically, LMGM fully leverages the linear complexity of Mamba. A three-directional scanning Mamba block is designed to exploit spatial correlations along the three dimensions of 3D seismic data. Meanwhile, a dual-domain-aware block is further developed to integrate time- and frequency-domain features for enhanced feature representation. In addition, a two-stage plug-and-play nonlinear normalization strategy is proposed to enhance the adaptability of LMGM to diverse seismic records while preserving signal characteristics. We employ 3D seismic data interpolation as a representative task to comprehensively evaluate LMGM. The results demonstrate that LMGM achieves superior processing performance while substantially reducing computational cost compared with U-Net-based and Transformer-based methods. Furthermore, experiments on random noise removal demonstrate the applicability of LMGM beyond interpolation, supporting its potential as a general GM framework.

physics.geo-ph↗

arXiv2026

From Gameplay to Policy: Towards Scalable Robot Data Collection via Gamified Robot-Free Interaction

Learning generalizable robot manipulation policies requires large-scale and diverse interaction data, yet collecting real-world demonstrations remains costly and difficult to scale. Existing approaches to data collection are either dependent on specific robot hardware that limits crowdsourcing and transferability, or suffer from incomplete annotation and limited behavioral diversity. Inspired by how games sustain long-term human engagement, we explore an alternative paradigm that turns data collection into an engaging gameplay experience and transfers the resulting human manipulation experience to real robots. We present Project Kitchen, a VR-based gamified egocentric data collection platform that elicits diverse, goal-directed manipulation while remaining independent of specific robot embodiments and hardware, making it applicable to broader and potentially large-scale deployment. To bridge the game-to-real gap, we further introduce Game2Policy, which extracts embodiment-invariant affordance cues, including contact points and sub-goal states, from gameplay trajectories. An affordance model is pre-trained on game-collected data and then jointly fine-tuned with downstream policies using only a handful of real-robot demonstrations. Experiments show that Game2Policy improves average success rates by 10.0 points in simulation and 18.3 points on real robots in the few-shot setting. User studies and quantitative analyses further show that Project Kitchen promotes diverse manipulation behaviors and provides an engaging data collection experience. These results demonstrate the potential of gamified virtual environments as a scalable source of manipulation knowledge. The platform and code will be released upon acceptance.

cs.RO↗

arXiv2026

LogiScope-VQA: Benchmarking Vision-Language Models for Logistics Hazard Identification in Industrial Scenarios

Large Multimodal Models (LMMs) large-scale deployment in industrial warehouse settings specifically necessitates that models exhibit human-expert-level hazard-oriented perception, understanding, and reasoning capabilities. However, the scarcity of real industrial data, tightly coupled to commercial terms, significantly hampers further advancement. To bridge this gap, we curate LogiScope-VQA to investigate the practical applicability of mainstream LMMs in real-world logistics operations. LogiScope-VQA comprises 2,476 images and 2,918 videos primarily sourced from real-world logistics parks, along with 10,274 VQAs meticulously curated and validated by human annotators. Grounded in 18 core objects and 20 risk types, we devise 39 subtasks aligned with three principal themes: industrial element perception, warehouse knowledge understanding, and potential risk reasoning. Furthermore, we incorporate dynamic thinking-budget configurations and dual-dimensional risk bias analyses to elucidate the properties of LMMs. Extensive experiments unveil that even powerful proprietary models, including GPT-5.5, Gemini-3.1-Pro, and Claude-Opus-4.7, exhibit a significant gap relative to human performance. The unique challenge of jointly integrating perception, understanding, and reasoning for hazard identification poses substantial headroom for further improvement on LogiScope-VQA. We additionally reveal the pervasive security bias issue that impedes LLMs' practical deployment in real-world settings. The industrial dataset is publicly available under the CC BY-NC-SA 4.0 license.

cs.CV↗

arXiv2026

Learning Terrain-Aware Whole-Body Control for Perceptive Legged Loco-Manipulation

Legged manipulators integrate exceptional terrain adaptability along with mobile manipulation capabilities, which make them highly promising for deployment in human-centric environments. By coordinating the control of both legs and arms, a whole-body controller can significantly expand the operational workspace of legged manipulators. However, many existing whole-body controllers primarily depend on proprioception and do not incorporate the critical exteroception required for effective terrain topology perception. This limitation can hinder their ability to adapt to varying environmental conditions and navigate complex terrains effectively. In this paper, we introduce TA-WBC, a terrain-aware whole-body control framework for legged manipulators, which features a novel RL-based unified policy tailored to whole-body loco-manipulation tasks in various terrains. Specifically, we employ a \rev{hierarchical exteroceptive encoder} to extract terrain features, providing an essential basis for the robot to proactively adapt posture and footholds. Furthermore, to facilitate stable cross-terrain loco-manipulation, we propose a novel end-effector sampling method based on the foot contact plane, \rev{decoupling the manipulation target from base height, roll, and pitch variations}. Moreover, a dual-policy distillation module is introduced to integrate expansive whole-body motion with terrain adaptability without catastrophic forgetting. The simulation and real-world experiments validate the robustness of our proposed controller, which leads to a larger reachable space, less tracking error, and reduced unexpected stumbles. This unified policy highlights the promising capabilities of legged manipulators in performing loco-manipulation tasks across complex terrains.

cs.RO↗

arXiv2026

Continuous Actions from Discrete Minds: Latent-Aligned Planning for End-to-End Autonomous Driving

Bridging the gap between the discrete reasoning of Vision-Language Models and the continuous, physics-constrained nature of autonomous driving remains a significant challenge. In this work, we introduce LaPla, a unified Vision-Language-Action (VLA) framework featuring latent-aligned planning to seamlessly ground semantic understanding in precise motion execution. We first design an action tokenizer based on a residual vector-quantized variational autoencoder (VQ-VAE), capturing vehicle kinematics and encoding trajectory features into a structured latent space. Rather than discrete codebook lookups that inevitably introduce quantization errors, LaPla repurposes this representation as a physical prior to bridge the modality gap between high-dimensional semantics and the raw action space. Specifically, given multimodal inputs integrating multi-view images, historical actions, and textual instructions, LaPla incorporates concurrent action queries to causally attend to the multimodal context in a single forward pass, projecting hidden states directly into the pretrained VQ-VAE latent space. The frozen decoder then translates these continuous latents into actions, effectively eliminating quantization errors and ensuring physically plausible trajectories while bypassing time-consuming autoregressive generation. Extensive experiments on the nuScenes benchmark demonstrate that LaPla achieves competitive open-loop performance, reducing long-horizon L2 error by 15.52% compared to state-of-the-art VLA methods. Closed-loop evaluations on the NVIDIA AlpaSim simulator further confirm its superior capability in ensuring smooth driving progress, improving the success rate by 33.34 percentage points with significantly reduced inference latency.

cs.CV↗

arXiv2026

Detecting Early and Late Divergences in Survival Curves Using Nonparametric Effect Measures

Clinical trials often show treatment curves that diverge early and converge later, or vice versa patterns that are poorly captured by the proportional-hazards assumption. We develop a joint inferential framework for two nonparametric functionals of censored survival data: the Kaplan--Meier-based Mann--Whitney effect and a novel temporal contrast separating early and late differences. The approach provides interpretable, probability-scale effect measures and enables joint inference for global and temporal contrasts under right censoring. In simulation studies, the method outperforms the log-rank test under non-proportional hazards while maintaining nominal type-I error. A real-world application illustrates how the temporal contrast reveals clinically meaningful early treatment advantages that remain hidden in standard analyses

stat.ME↗

arXiv2026

A Frequency-Aware Dynamic Knowledge Distillation Framework: An Effective Tool for Bridging Low- and High-Frequency Seismic Information

Seismic data contain rich information across different frequency bands, with low-frequency components primarily characterizing large-scale geological structures and high-frequency components preserving fine-scale seismic details. Effectively integrating these frequency-dependent components is essential for seismic feature learning to better preserve structural continuity and fine-scale details. Knowledge distillation provides an effective means for transferring informative representations from high-quality data. However, existing distillation-based frameworks usually treat seismic features in a full-band manner, ignoring relationships across frequency bands and thereby limiting the coordinated transfer of low- and high-frequency knowledge. To bridge low- and high-frequency seismic features through knowledge distillation, we propose a frequency-aware dynamic knowledge distillation framework (FADKD-Net), which establishes a teacher-student learning framework and performs frequency-aware knowledge transfer between low- and high-frequency bands. Specifically, FADKD-Net decomposes seismic features into low- and high-frequency components and performs targeted distillation to exploit their complementary information. Low-frequency distillation guides the student model to learn stable structural priors, thereby improving the overall continuity of seismic events. Meanwhile, high-frequency distillation enhances detailed feature modeling and improves the representational capability for complex and small-scale structures. Furthermore, a cross-domain feature alignment strategy is proposed to reduce distributional discrepancies across different surveys and enhance the transferability of the seismic representations learned by FADKD-Net.

physics.geo-ph↗

arXiv2026

Evolution of Stellar Activity and Habitable Zone (EATEN): III. X-ray Activity of Dwarfs in Open Clusters and Field Stars

Stellar X-ray emission serves as a direct diagnostic of coronal activity, which is fundamentally linked to coronal heating processes. It also strongly influences the atmospheres and long-term habitability of orbiting exoplanets. Investigating how this high-energy emission evolves is therefore essential for understanding the evolution of stellar magnetic dynamos and planetary atmospheres and habitability. In this work, we investigate the evolution of X-ray activity and XUV irradiation for a sample of F-M dwarf stars based on Chandra and XMM-Newton observations. We find that F- and G-type stars broadly follow the traditional evolutionary picture of an early saturated (or weakly declining) phase followed by a modest decline, whereas K- and M-type stars exhibit a clear three-phase evolution of a saturated phase, an intermediate phase of rapid decay, and a final modest decline phase. By combining X-ray, ultraviolet, and Ca II H&K bands, we show that coronal emission becomes increasingly dominant toward lower-mass stars. Based on the cumulative XUV emission calculated from our fitted relation, planets around F- and G-type stars experience relatively moderate XUV environments, while those around K- and M-type stars may exceed the empirical cosmic shoreline shortly after reaching the main sequence, though this conclusion depends on the adopted shoreline value.

astro-ph.SR↗

arXiv2026

Behavior Specification-Guided Program Synthesis for Binary Deobfuscation

Deobfuscation is critical to reverse engineering and security analysis because it restores the readability and analyzability of obfuscated code. However, existing research primarily focuses on source-code deobfuscation, while binary-level deobfuscation remains largely underexplored despite its practical importance when source code is unavailable. Existing binary deobfuscation methods typically decompile binaries into pseudocode and then apply structural transformations. However, because compilation discards high-level semantics such as precise type information and source-level structures, this decompilation-based paradigm often produces low-quality code and provides limited assurance that the recovered code preserves the runtime behavior of the original program. To address these limitations, we propose a paradigm shift from structural transformation to behavior-driven synthesis. Our core insight is that although obfuscation distorts a program's internal structure, semantics-preserving transformations must retain its observable execution behavior. Based on this insight, we introduce BinMirror, an approach that reformulates binary deobfuscation as a behavior-specification-guided program synthesis task. By treating dynamic execution traces and interaction snapshots as behavioral specifications, BinMirror synthesizes high-quality source code and validates it against runtime observations collected from heavily obfuscated binaries. Extensive evaluations on 1.5 million synthetically obfuscated binaries show that BinMirror significantly outperforms state-of-the-art baselines, achieving a unit-test Pass@1 of 74.5% under extreme obfuscation. These results demonstrate the practical utility of BinMirror in restoring semantic clarity for real-world security analysis.

cs.SE↗

arXiv2026

Tac4Loco: Learning Spatiotemporal Plantar Pressure Representations for Humanoid Locomotion

Humanoid robots are expected to traverse complex terrains, where the plantar support may vary dramatically due to foot placement errors, ground properties, and transient dynamics. To achieve robust locomotion, the robots are required to adapt to uneven terrain and uncertain foot--ground interactions. Existing locomotion policies rely primarily on proprioception or exteroceptive terrain perception, where the former provides only indirect evidence of plantar support, while the latter predicts contact conditions before touchdown but cannot observe the actual support in real-time. Although some studies incorporate plantar contacts as an auxiliary perception, they rely mainly on summary statistics, overlooking the spatial topology of plantar pressure, which provides a more direct characterization of the realized contact state. To bridge this gap, we present Tac4Loco, a tactile-perceptive framework that incorporates multi-array plantar pressure as direct feedback for humanoid locomotion. We formulate a topology-preserving ordinal representation to map simulated and physical sensor signals into a shared observation space, with a dual-branch encoder for extracting their spatial and temporal representations. Subsequently, the learned spatiotemporal features are integrated with augmented proprioception including terrain estimation cues, and provided to an asymmetric actor-critic architecture for policy learning. Extensive simulation and real-world experiments demonstrate improved tracking performance and support adaptation on terrains with inclined, partial, asymmetric, and changing support. We further demonstrate its zero-shot deployment on unseen compliant and unstructured terrains, including a foam platform and a gravel road. All code and experimental configurations will be released as open-source to facilitate reproducibility.

cs.RO↗