CV

Research Intern

Embodied Agent Group, ModelBest Inc.

Advised by Zhiyuan Liu; latent action prediction in vision-language models.

• Investigated state-of-the-art multimodal and video-based embodied learning methods, including Flamingo, Genie, UniPi, Video PreTraining (VPT), and AVDC, and explored latent-action learning by predicting future visual states and recovering control actions from frame transitions via inverse dynamics.

Research Intern

Intelligent Control Lab, Carnegie Mellon University

Advised by Changliu Liu; human-robot collaboration and time-series prediction.

• Proposed and developed a multimodal hierarchical human-robot collaboration framework integrating vision, speech, hierarchical plan prediction, and online user adaptation; deployed it on a Kinova Gen3 for long-horizon toy car assembly, achieving 91.8% task success, 15.9% faster completion, and 84% user satisfaction.

Undergraduate Research Intern

Computer Vision Lab, Tsinghua University

Advised by Jiwen Lu; 2D image super-resolution.

• Co-developed Implicit Pixel Flow (IPF), a model-agnostic coordinate-warping framework for photorealistic arbitrary-scale super-resolution.

Ph.D. Candidate

Carnegie Mellon University

Electrical and Computer Engineering

• Expected graduation in May 2029.
• Courses: Introduction to Robot Learning (16-831), Advanced Control for Robotics (16-714), Kinematics for Robot Manipulation (16-741), Estimation, Detection, and Learning (18-752), Safe Robotics (16-883).
• Advised by Changliu Liu.

M.S.

Carnegie Mellon University

Electrical and Computer Engineering

B.Eng.

Tsinghua University

Automation

ManiSkillFormer: Demonstration-Free Compositional Manipulation via Geometric Contracts and Agentic Skill Graph

Submitted to ICRA 2027

Peiqi Yu, Mosam Dabhi, Shangtao Li, Bowei Li, Laszlo Jeni, Changliu Liu

Website

Emergent Neural Automaton Policies: Learning Symbolic Structure from Visuomotor Trajectories

Robotics: Science and Systems (RSS)

Yiyuan Pan*, Xusheng Luo*, Hanjiang Hu, Peiqi Yu, Changliu Liu

Website

Autonomous Integration and Improvement of Robotic Assembly using Skill Graph Representations

ICRA Workshop on Multi-Agent Robotic Systems: Real-World Collaboration and Interaction

Peiqi Yu*, Philip Huang*, Chaitanya Chawla*, Guanya Shi, Jiaoyang Li, Changliu Liu

Website

Dynamic Neural Koopman Distillation for Real-Time Robot Control Using Diffusion Models

Preprint

Lei Zheng, Peiqi Yu, Zengqi Peng, Changliu Liu, Armin Lederer

Website

NeSyPack: A Neuro-Symbolic Framework for Bimanual Logistics Packing

RSS Workshop on Benchmarking Robot Manipulation

Bowei Li*, Peiqi Yu*, Zhenran Tang, Han Zhou, Yifan Sun, Ruixuan Liu, Changliu Liu

Website

Robustifying Long-term Human-Robot Collaboration through a Multimodal and Hierarchical Framework

IFAC Cyber-Physical & Human Systems

Peiqi Yu*, Abulikemu Abuduweili*, Ruixuan Liu, Changliu Liu

Recovering Realistic Details for Magnification-Arbitrary Image Super-Resolution

IEEE Transactions on Image Processing

Cheng Ma, Peiqi Yu, Jiwen Lu, Jie Zhou

First Prize, ICRA WBCD Competition

IEEE International Conference on Robotics and Automation

Second Prize, Electronics Design Contest

Tsinghua University

Third Prize, Challenge Cup

Tsinghua University