I am a 5th year Ph.D student in Assured Reliable Interactive Autonomous (ARIA) Systems Group and in Programming Languages and Verification at the University of Colorado Boulder (CUPLV ) at the University of Colorado Boulder.

Current Work

My ultimate goal is to easily teach robots the task specifications through demonstrations and execute the task autonomously. Especially, my work focuses on Learning from Demonstrations (LfD), i.e., the robot learns the specifications of a task through demonstrations (in the form of Formal Language, e.g., LTL). I am interested in learning a high-level abstract task as a formal language so that robots can utilize the learned specifications to adapt to similar environments.

Interests

My research focuses on task learning and strategy synthesis for robots. Especially, I am interested in the intersection between Formal Method, Optimization, and Robotics. I also worked intensively on Control Theory and Reinforcement Learning algorithms during my master’s degree (LQR, MPC, H-Infinity, Robust Control, Hierarchical RL, Model-based RL, etc.)

Bio

I completed my Bachelor and Master at Keio University in Japan, under the supervision of Masaki Takahashi. I also studied at the University of Illinois at Urbana-Champain for a year. Majored in Aerospace Engineering and Computer Science.

Short Videos

Optimal Planning for Timed Partial Order Specifications

Probabilistic Specification Learning for Planning with Safety Constraints