Konwoo Kim

About

Konwoo Kim is a rising researcher in the field of safe and generalizable robot learning, with a focus on constraint inference and human-robot interaction. His most-cited work, "Learning Shared Safety Constraints from Multi-task Demonstrations" (2023, 3 citations), introduces a novel framework for inferring implicit, task-agnostic safety rules from expert demonstrations—such as a kitchen robot learning that plates must never be broken, regardless of whether it is making a sandwich or clearing the table. This contribution addresses a critical gap in reinforcement learning: the challenge of manually specifying safety constraints for every possible task. By enabling robots to autonomously deduce shared behavioral boundaries across diverse tasks, Kim’s approach enhances both the safety and transferability of learned policies. His work is particularly impactful for real-world deployment, where robots must operate reliably in unstructured environments. Though early in his career, Kim’s research has already been recognized for its potential to bridge the gap between task-specific learning and generalizable safety, marking him as a promising voice in the growing field of constrained reinforcement learning and human-aligned AI.

Research Focus

Key Achievements

Top Papers

1. 1
   Learning Shared Safety Constraints from Multi-task Demonstrations

   3 citations · 2023

Key Collaborators

Contact & Links