Pengxiang Ding
Scholar

Pengxiang Ding

Google Scholar ID: QyBSTzEAAAAJ
Zhejiang University
Human Motion PredictionLarge Language ModelEmbodied AI
Citations & Impact
All-time
Citations
552
 
H-index
13
 
i10-index
16
 
Publications
20
 
Co-authors
14
list available
Resume (English only)
Academic Achievements
  • Published several papers, notable works include:
  • - Long-VLA: Unleashing Long-Horizon Capability of Vision Language Action Model for Robot Manipulation
  • - CARP: Visuomotor Policy Learning via Coarse-to-Fine Autoregressive Prediction
  • - CEED-VLA: Consistency Vision-Language-Action Model with Early-Exit Decoding
  • - SSR: Enhancing Depth Perception in Vision-Language Models via Rationale-Guided Spatial Reasoning
  • - Unicorn: Text-Only Data Synthesis for Vision Language Model Training
  • - Openhelix: A short survey, empirical analysis, and open-source dual-system vla model for robotic manipulation
  • - Humanoid-vla: Towards universal humanoid control with visual integration
  • - Accelerating vision-language-action model integrated with action chunking via parallel decoding
  • - Rethinking Latent Representations in Behavior Cloning: An Information Bottleneck Approach for Robot Manipulation
  • - QUART-Online: Latency-Free Large Multimodal Language Model for Quadruped Robot Learning
  • - GEVRM: Goal-Expressive Video Generation Model For Robust Visual Manipulation
Research Experience
  • Current research focuses on embodied AI, particularly in the VLA direction. Served as the principal investigator or co-investigator in multiple research projects.
Education
  • Currently a third-year Ph.D. student at Zhejiang University, advised by Prof. Donglin Wang. Also involved in a joint program with Westlake University as a member of Machine Intelligence Laboratory (MiLAB). Prior to his Ph.D., he received his Msc. Degree from the School of Artificial Intelligence, Beijing University of Posts and Telecommunications in 2022, advised by Prof. Jianqin Yin.
Background
  • Research Interests: Embodied AI, including VLA/VLM/World Model. Mainly focused on the VLA direction. Published 15 papers as first author, co-first author, or project leader.
Miscellany
  • Personal interests and hobbies not mentioned