Yuki (Huaxiaoyue) Wang

PoRTaL lab member. 3rd Year CS Ph.D. at Cornell University.

yuki_profile_pic.jpg

yukiwang at cs.cornell.edu

hw575 at cs.cornell.edu

Hi! I am a third-year CS PhD student at Cornell Univeristy. I am currently a member of the PoRTaL lab, working with Prof. Sanjiban Choudhury on how foundation models can help robots learn from interacting with humans.

Foundation models are used in many ways, but I am particularly interested in how they can help robots learn from humans!

When humans learn to do tasks, we can watch others, ask questions, and improve based on feedback instead of simply relying on language instructions. My key insight is: robots should do the same; they can also benefit from (1) demonstrations, (2) online interactions, and (3) feedback when they need to learn new tasks and refine their behaviors.

I did my undergrad at Harvey Mudd College, where I did research in a wide range of areas: experimental airplane construction (@HMC Aero Lab), underwater robotics (@LAIR), and human-robot interactions(@HEATlab).

Link to my CV: here

news

Nov 4, 2024 Two first-authored papers accepted to CoRL 2024! APRICOT presents a novel algorithm that allows LLMs to systematically query users to reduce uncertainty, and MOSAIC showcases a multi-robot system that collaborates with users to cook!
May 17, 2024 My first co-authored paper MOSAIC with Kushal Kedia and Juntao Ren won the best paper award at the VLNMN workshop and the best poster at the MoMa workshop at ICRA 2024!
Oct 2, 2023 My first paper Demo2Code got accepted to Neurips 2023!

selected publications

  1. APRICOT: Active Preference Learning and Constraint-Aware Task Planning with LLMs
    Huaxiaoyue Wang, Nathaniel Chin, Gonzalo Gonzalez-Pumariega, and 5 more authors
    2024
  2. MOSAIC: Modular Foundation Models for Assistive and Interactive Cooking
    Huaxiaoyue Wang, Kushal Kedia, Juntao Ren, and 14 more authors
    2024
  3. Time Your Rewards: Learning Temporally Consistent Rewards from a Single Video Demonstration
    Huaxiaoyue Wang, William Huey, Anne Wu, and 2 more authors
    2024
  4. Demo2Code: From Summarizing Demonstrations to Synthesizing Code via Extended Chain-of-Thought
    Huaxiaoyue Wang, Gonzalo Gonzalez-Pumariega, Yash Sharma, and 1 more author
    2023