Luke Shimanuki

Luke  Shimanuki
MIT EECS | Hudson River Trading Undergraduate Research and Innovation Scholar
Advisor: Leslie P. Kaelbling
Department: EECS
Years: 2018-2019
Research Project Title:

Learning Hierarchies in POMDPs

abstract:In the last half century, robots have proliferated across all sectors of industry, replacing humans and automating increasingly complex tasks. However, the field still struggles to move beyond single-purpose robots hand-engineered to accomplish a predefined task in a particular domain. My research focuses on developing algorithms that enable robots to learn from experience and observations and generalize their understanding to new and complex scenarios. I will start by working on models to estimate the difficulty of potential subtasks, then I will integrate this with planning algorithms to allow the robot to re-sequence learned subtasks to solve new problems.

"I am participating in SuperUROP because I want to continue to build on my prior research in robotics and learning and eventually determine a path forward for graduate school and beyond."