Novel Reinforcement Learning Methods in Collaborative Environments
CSEF · 2019 Mathematical Sciences
Overview
Objectives Reinforcement learning (RL) is an emerging field with numerous applications in a wide variety of domains, such as robotics, supply chain and network optimization, marketing, and more. Our research focuses on developing improved RL algorithms and applying them to domains involving complex interactions between multiple intelligent systems. Methods We propose a novel algorithm which employs a Bayesian method of parameter space exploration to solve reinforcement learning problems. We parameterize policies as neural networks, and a Gaussian process is used to learn the expected return of a policy given the parameters. The system is trained by updating the parameters in the directions suggested by the Gaussian Process to maximize the expected return and explore important new states. Results Our method is applied to three challenging robotic simulations where we achieve gains of 8%, 20%, and 33% over current methods. We also observe good performance in cooperative and competitive environments that require communication between multiple intelligent systems. Conclusions The improved performance of our methods will enable the development of intelligent systems that tackle significantly more complex problems in robotics and other fields. Our novel method of Bayesian parameter space exploration can be theoretically shown to result in less conservative updates towards better policies.
Summary statement
We develop a novel method of Bayesian parameter space exploration inspired by Policy Gradient methods to solve Reinforcement Learning Problems.
Help received
We designed, developed, and implemented our algorithm by ourselves, but received some help. In particular, we emailed Dr. Yi Wu (UC Berkeley) and Dr. Ryan Lowe (McGill University) to clarify some questions we had on their papers.
Competition history
- CSEF 2019
Resources
Related projects
ISEF · 2019
Novel Reinforcement Learning Methods in Collaborative Environments
CSEF · 2017
Coordinated Multi-Agent Control Utilizing Deep Reinforcement Learning
ISEF · 2018
Optimizing Reinforcement Learning Through Dynamic Environment Manipulation
ISEF · 2021
Assessing the Impact of Robust Stochastic Operators on Q-learning Efficiency in an OpenAI Gym Environment with a Large Observation Space
ISEF · 2021
Optimizing Human Feedback in Reward Modelling
CSEF · 2017
Direct Control of Autonomous Mobile Robots via Deep Reinforcement Learning
CSEF · 2011
A Study and Analysis of Q-Learning Algorithm Parameters for Decision Making using a Developed Simulation Tool
ISEF · 2026
Meta-Reinforcement Learning Framework for Network-Level Bridge Management
Closest projects by meaning, across every fair and year in the corpus.
Browse more like this
Source: California Science & Engineering Fair public projects