H
Hoang M. Le
Researcher at California Institute of Technology
Publications - 31
Citations - 833
Hoang M. Le is an academic researcher from California Institute of Technology. The author has contributed to research in topics: Reinforcement learning & Imitation. The author has an hindex of 13, co-authored 23 publications receiving 642 citations. Previous affiliations of Hoang M. Le include Bucknell University & Microsoft.
Papers
More filters
Posted Content
Coordinated Multi-Agent Imitation Learning
TL;DR: It is shown that having a coordination model to infer the roles of players yields substantially improved imitation loss compared to conventional baselines, and the method integrates unsupervised structure learning with conventional imitation learning.
Proceedings Article
Batch Policy Learning under Constraints
TL;DR: In this paper, the problem of batch policy learning under multiple constraints is studied, and a flexible meta-algorithm that admits any batch reinforcement learning and online learning procedure as subroutines is proposed.
Posted Content
Empirical Study of Off-Policy Policy Evaluation for Reinforcement Learning
TL;DR: This work presents the first comprehensive empirical analysis of a broad suite of OPE methods, and offers a summarized set of guidelines for effectively using OPE in practice, and suggest directions for future research.
Data-driven ghosting using deep imitation learning
TL;DR: In this paper, a data-driven ghosting method using advanced machine learning methodologies called "deep imitation learning" was applied to a season's worth of tracking data from a recent professional league in soccer, which enables counterfactual analysis of effectiveness of defensive positioning as both a measurable and viewable quantity.
Proceedings Article
Coordinated multi-agent imitation learning
TL;DR: The authors propose a joint approach that simultaneously learns a latent coordination model along with the individual policies for fine-grained behavior modeling in team sports, where different players occupy different roles in the coordinated team strategy, and show that having a coordination model to infer the roles of players yields substantially improved imitation loss compared to conventional baselines.