Haibo Zhao

I'm a research assistant in Helping Hands Lab at Northeastern University, where I am fortunate to be advised by Prof. Rob Platt and Prof. Robin Walters.

I'm working on robot manipulation in unstructured environments. I'm passionate about solving challenging problems with high impact.

Email  /  CV  /  Scholar  /  X  

Research

We present an equivariant diffusion policy approach for sample-efficient robotic learning that leverages symmetries in robotic manipulation tasks.

We introduce a hierarchical skill learning framework that leverages vlm/mllm agent as high level, diffusion-based policy as low level, and object-centric representation to generalize across spatial arrangements, object instances, and novel tasks using minimal demonstrations.

We introduce BEAR, a large-scale benchmark and evaluation framework for assessing multimodal large language models on step-wise embodied capabilities. BEAR spans perception, reasoning, and action understanding across 4,469 interleaved image–video–text samples, providing a comprehensive analysis of atomic embodied skills.

We introduce a tactile equivariance framework for residual rotation correction in robotic manipulation tasks.

By introducing a frame transfer interface, we impose a soft constraint from the high-level open-loop policy onto the low-level closed-loop policy, combining the strengths of both approaches.

Service

• Reviewer for CoRL 2025
• Reviewer for RA-L 2025