Pablo Hernandez Leal

Photo of Pablo Hernandez Leal

Researcher

PhD Computer Science, Instituto Nacional de Astrofísica, Optica y Electronica

Pablo is a researcher working in the group led by Matt Taylor. Pablo is interested in how learning algorithms developed for single-agent environments should be adapted to multiagent settings. One of his objectives is to propose efficient multiagent learning algorithms for strategic interactions using models and concepts from game theory, Bayesian reasoning, and reinforcement learning. 

Before joining Borealis Pablo studied at INAOE in Mexico and at Washington State University in the USA, later he worked as a researcher at CWI, the National Research Institute for Mathematics and Computer Science of the Netherlands.

Born and raised in Mexico, Pablo loves tacos and spicy food. After living in Amsterdam for a couple of years, he likes biking to work although the Canadian weather sometimes makes this impossible.

pablo.hernandez@borealisai.com

Research Areas

Reinforcement Learning

Publications

May 9, 2020

Temporally Extended Auxiliary Tasks

Workshop on Adaptive and Learning Agents (AAMAS), 2020
Authors: C. Sherstan, B. Kartal, P. Hernandez-Leal , M. E. Taylor
Feb. 7, 2020

Uncertainty-Aware Action Advising for Deep Reinforcement Learning Agents

Thirty-Fourth AAAI Conference on Artificial Intelligence, 2020
Authors: F. L. Da Silva, P. Hernandez-Leal , B. Kartal, M. E. Taylor
Nov. 10, 2019

Towers of Saliency: A Reinforcement Learning Visualization Using Immersive Environments

ACM Interactive Surfaces and Spaces (ISS), 2019
Authors: N. Douglas, D. Yim, B. Kartal, P. Hernandez-Leal , M. E. Taylor, F. Maurer
Oct. 8, 2019

Agent Modeling as Auxiliary Task for Deep Reinforcement Learning

AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), 2019
Authors: *B. Kartal, *P. Hernandez-Leal , M. E. Taylor
* Denotes equal contribution
Oct. 8, 2019

Terminal Prediction as an Auxiliary Task for Deep Reinforcement Learning

AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), 2019
Authors: *B. Kartal, *P. Hernandez-Leal , M. E. Taylor
* Denotes equal contribution
Oct. 8, 2019

Action Guidance with MCTS for Deep Reinforcement Learning

AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), 2019
Authors: *B. Kartal, *P. Hernandez-Leal , M. E. Taylor
* Denotes equal contribution
Oct. 8, 2019

On Hard Exploration for Reinforcement Learning: a Case Study in Pommerman

AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE), 2019
Authors: C. Gao, B. Kartal, P. Hernandez-Leal , M. E. Taylor
July 7, 2019

Skynet: A Top Deep RL Agent in the Inaugural Pommerman Team Competition

The Multidisciplinary Conference on Reinforcement Learning and Decision Making (RLDM), 2019
Authors: C. Gao, P. Hernandez-Leal , B. Kartal, M. E. Taylor
May 13, 2019

Safer Deep RL with Shallow MCTS: A Case Study in Pommerman

Workshop on Adaptive Learning Agents (AAMAS), 2019
Authors: B. Kartal, P. Hernandez-Leal , C. Gao, M. E. Taylor
May 13, 2019

A Survey and Critique of Multiagent Deep Reinforcement Learning

Journal of Autonomous Agents and Multiagent Systems (JAAMAS), 2019
Authors: P. Hernandez-Leal , B. Kartal, M. E. Taylor
Jan. 27, 2019

Using Monte Carlo Tree Search as a Demonstrator within Asynchronous Deep RL

Workshop on Reinforcement Learning in Games (AAAI), 2019
Authors: B. Kartal, P. Hernandez-Leal , M. E. Taylor
Dec. 3, 2018

Skill Reuse in Partially Observable Multiagent Environments

Workshop on Latinx in AI Coalition (NeurIPS), 2018
Authors: P. Hernandez-Leal , B. Kartal, M. E. Taylor
Picture of Pablo Hernandez Leal