Haode Zhang 张皓德

I am a first-year Master's student at Carnegie Mellon University. I received my Bachelor's degree from Shanghai Jiao Tong University. I was fortunate to work with Prof. Ruohan Gao and Prof. Wanxin Jin during my undergraduate.

Email  /  Scholar  /  Twitter  /  LinkedIn

Publications

(* indicates equal contribution; ^† indicates equal advising)
Highlighted entries are representative works.

We propose OmniTacTune, a two-stage residual RL framework that efficiently adapts tactile feedback to pretrained visual policies. It requires neither offline tactile demonstrations nor base-policy fine-tuning and generalizes across diverse contact-rich tasks, visual policies, and tactile representations. Across four real-world tasks, it improves visual base policies from 5–40% to 85–100% success within only 40–80 minutes of online interaction.

We propose a novel geometric view of reward alignment as an iterative cutting process over the hypothesis space. Our batched cutting method significantly improves data efficiency by maximizing the value of each human preference query. We introduce a conservative cutting algorithm that ensures robustness to unknown erroneous human preferences without explicitly identifying them.

We propose Scene Dynamic Field (SDF), a cost-efficient framework that integrates physics simulators into multi-task fine-tuning, substantially improving MLLMs’ intuitive physics understanding and achieving strong generalization across physical domains.

Service

Conference Reviewer: ECCV 2026, CoRL 2026

The website is built upon this template.