Sitemap
A list of all the posts and pages found on the site. For you robots out there, there is an XML version available for digesting as well.
Pages
Posts
portfolio
Upper-Limb Motion Recognition
Published:
Early recognition of upper-limb motions for exoskeleton-assisted mirror rehabilitation.
Practical Vision-Language-Action Models
Published:
Benchmarking and improving the practical capabilities of vision-language-action models.
DySta
Published:
Static-dynamic disentanglement for efficient multi-frame vision-language-action models.
WAM and Robot Learning
Published:
Robot learning projects involving WAM pre-training, robot data processing, model training, and evaluation.
RoboMemArena
Published:
A keyframe-annotated benchmark for memory-dependent and long-horizon robotic manipulation, together with the PrediMem method.
publications
Exploring Disentangled Appearance-Motion Contexts for Temporal Activity Localization
Published in International Joint Conference on Neural Networks (IJCNN), 2025, 2025
Explores disentangled appearance-motion contexts for temporal activity localization.
Manipulating the Bounding Box: Multimodal Controlled Backdoor Attacks on 3D Visual Grounding Models
Published in International Joint Conference on Neural Networks (IJCNN), 2025, 2025
Multimodal controlled backdoor attacks against 3D visual grounding models.
MonoAttack: A Strong Attack Framework with Depth-Migration and Attribute-Tampering for Monocular 3D Object Detection
Published in International Joint Conference on Neural Networks (IJCNN), 2025, 2025
An attack framework for monocular 3D object detection based on depth migration and attribute tampering.
High-Accuracy Early Recognition of Upper-Limb Motions for Exoskeleton-Assisted Mirror Rehabilitation
Published in IEEE Robotics and Automation Letters (RA-L), 2025, 2025
Early upper-limb motion recognition for exoskeleton-assisted mirror rehabilitation.
Rethinking the Practicality of Vision-Language-Action Model: A Comprehensive Benchmark and An Improved Baseline
Published in IEEE International Conference on Robotics and Automation (ICRA), 2026, 2026
A comprehensive benchmark and improved baseline for evaluating the practical capabilities of vision-language-action models.
HCSG: Human-Centric Semantic-Geometric Reasoning for Vision-Language Navigation
Published in Preprint, 2026, 2026
A human-centric semantic-geometric reasoning framework for vision-language navigation.
EnvTrap: Revealing the Environment-Only Attack Surface in Embodied AI via Consequence-Blind Action Execution
Published in NeurIPS 2026, Poster, 2026
Analyzes an environment-only attack surface in embodied AI caused by consequence-blind action execution.
Static-Dynamic Disentanglement for Efficient Multi-Frame Vision-Language-Action Models
Published in NeurIPS 2026, Poster, 2026
Disentangles static and dynamic visual information to improve the efficiency of multi-frame vision-language-action models.
RoboMemArena: A Keyframe-Annotated Benchmark for Memory-Dependent Robotic Manipulation
Published in Preprint, 2026, 2026
A keyframe-annotated benchmark for evaluating memory-dependent and long-horizon robotic manipulation.
EmbodiedSmith: Scaling Embodied Data through Recursive Self-Improvement Flywheel in Simulation
Published in Preprint, 2026, 2026
A simulation framework for scalable embodied-data generation that jointly refines assets, scenes, and tasks through recursive self-improvement.