A Narration-based Reward Shaping Approach using Grounded Natural Language Commands

no code implementations • 31 Oct 2019 • Nicholas Waytowich, Sean L. Barton, Vernon Lawhern, Garrett Warnell

While this problem can be addressed through reward shaping, such approaches typically require a human expert with specialized knowledge.

Starcraft Starcraft II

Grounding Natural Language Commands to StarCraft II Game States for Narration-Guided Reinforcement Learning

no code implementations • 24 Apr 2019 • Nicholas Waytowich, Sean L. Barton, Vernon Lawhern, Ethan Stump, Garrett Warnell

While deep reinforcement learning techniques have led to agents that are successfully able to learn to perform a number of tasks that had been previously unlearnable, these techniques are still susceptible to the longstanding problem of {\em reward sparsity}.

Reinforcement Learning (RL) Starcraft +1

Coordination-driven learning in multi-agent problem spaces

no code implementations • 13 Sep 2018 • Sean L. Barton, Nicholas R. Waytowich, Derrik E. Asher

We discuss the role of coordination as a direct learning objective in multi-agent reinforcement learning (MARL) domains.

Multi-agent Reinforcement Learning reinforcement-learning +1