Takuma Yoneda

Takuma Yoneda

  • United States
  • PhD candidate at TTIC
  • Reinforcement Learning, Robotics and NLP
  • takuma (at) ttic.edu

I am a PhD candidate at TTIC, an academic institute located on the campus of the University of Chicago.
My research interest lies in the intersection of Reinforcement Learning and Natural Language Processing, especially in the context of robotics.

In the past, I have interned at several places including UCL (2018 summer) and Preferred Networks (2019 summer) which is known for Chainer.

I am now at Robot Intelligence through Perception Lab working with Ge Yang, Bradly Stadie and Matthew R. Walter.
TTIC featured me recently: Takuma Yoneda - Student Highlight

Publications (Google Scholar)




Dexterious Manipulation with a three-finger robot

We participated in Real Robot Challenge, where participants develop a method that can grasp and manipulate an object with TriFinger robot. We won 1st place in both competition phases (Phase 2 & 3), and received 12,500 EUR as a prize.
[code] [paper] [videos]

Auxiliary Tasks for Navigation in Unknown Environments (class project)

We used Habitat simulator to train an agent that navigates itself in a photo-realistic 3-D indoor environment. The agent only observes first-person view image. We trained the agent with several different auxiliary tasks such as depth prediction, and compare them to see what auxiliary tasks can help navigation. The gif on the left shows a PPO agent trained with reward (distance to the goal) and auxiliary depth-prediction.