wil thomason

I am a fifth-year PhD candidate in the Department of Computer Science at Cornell University. I work with Professor Hadas Kress-Gazit on combining symbolic and geometric planning. Specifically, I’m interested in improving the efficiency, robustness, and utility of robot task and motion planning.

Though I primarily research robot planning, I’m also interested in ML and formal methods for robotics, robotics research broadly, and programming languages. I am graciously funded by the NDSEG Fellowship.

Research

I work on combining symbolic and geometric planning, often referred to as “integrated task and motion planning”. My work explores approaches to task and motion planning that provide high performance and robustness to action failure with lightweight domain-specific assumptions. I also research approaches to improved planning effort reuse through abstract plan representation and fusions of classical planning with modern machine learning for more efficient, automatic planning and execution.

For more about my research, please see my publication and presentation list.

News

Teaching

Outreach

  • I was a mentor for Black in AI, providing assistance and advice to my mentee on his Ph.D. program application process. (Fall 2019 - Spring 2020)
  • I have helped review extended abstracts for the Black in AI workshop. (Fall 2017, 2018, and 2019)
  • I have taught 7th and 8th grade girls about state machines and programming at Cornell’s Expanding your Horizons conference. (Spring 2016, 2017, and 2018)

Departmental Service