A Scalable Finite Difference Method for Deep Reinforcement Learning

no code implementations • 14 Oct 2022 • Matthew Allen, John Raisbeck, Hakho Lee

Several low-bandwidth distributable black-box optimization algorithms in the family of finite differences such as Evolution Strategies have recently been shown to perform nearly as well as tailored Reinforcement Learning methods in some Reinforcement Learning domains.

reinforcement-learning Reinforcement Learning (RL)

Provenance-Based Interpretation of Multi-Agent Information Analysis

no code implementations • 8 Nov 2020 • Scott Friedman, Jeff Rye, David LaVergne, Dan Thomsen, Matthew Allen, Kyle Tunis

Analytic software tools and workflows are increasing in capability, complexity, number, and scale, and the integrity of our workflows is as important as ever.

Evolution Strategies Converges to Finite Differences

no code implementations • 27 Dec 2019 • John C. Raisbeck, Matthew Allen, Ralph Weissleder, Hyungsoon Im, Hakho Lee

Since the debut of Evolution Strategies (ES) as a tool for Reinforcement Learning by Salimans et al. 2017, there has been interest in determining the exact relationship between the Evolution Strategies gradient and the gradient of a similar class of algorithms, Finite Differences (FD).

reinforcement-learning Reinforcement Learning (RL)