Lechen Zhang

I'm a first year Robotics PhD student at MBZUAI. Prior to that, I was a graduate research assistant in the Creative Machines Lab at Columbia University under Professor Hod Lipson.

Email / Scholar / Github / LinkedIn

Research

I'm interested in the intersection of computer vision, deep learning, generative AI, and robotics. My research focuses on understanding the physical world — including shape, motion, depth, and appearance — from sensor observations, with applications in robotic design and human-robot interaction.

	Eye, Robot: Learning Realistic Robot Gaze From Human Motion Data Under Review Realistic Gaze Transformer (RGT), a Transformer-VAE framework that learns full-head gaze dynamics, including head rotations and eyelid movements, from human motion capture data.
	AutoURDF: Unsupervised Robot Modeling from 4D Point Cloud Jiong Lin, Lechen Zhang, Kwansoo Lee, Jialong Ning, Judah Goldfeder, Hod Lipson CVPR 2025 (Acceptance Rate: 22.1%) project page / arXiv An unsupervised approach for understanding robot motion and constructing description files for unseen robots from point cloud frames.
	MoD-SLAM: Monocular Dense Mapping for Unbounded 3D Scene Reconstruction Heng Zhou, Zhetao Guo, Yuxiang Ren, Shuhong Liu, Lechen Zhang, Kaidi Zhang, Mingrui Li, IEEE Robotics and Automation Letters (RA-L) arXiv Monocular SLAM with metric depth estimation and Gaussian-based unbounded scene representation.
	Soft Robot Neural Evolution with LLMs Supervision Lechen Zhang ICRA 2024, Workshop on Co-design in Robotics, Oral project page / arXiv A computational framework for automatically designing soft robot morphologies using large language models for design guidance and physics simulation for evaluation.

The website template is from source code by Jon Barron.