A Comparison of learning algorithms on the Arcade Learning Environment

10/31/2014
by   Aaron Defazio, et al.
0

Reinforcement learning agents have traditionally been evaluated on small toy problems. With advances in computing power and the advent of the Arcade Learning Environment, it is now possible to evaluate algorithms on diverse and difficult problems within a consistent framework. We discuss some challenges posed by the arcade learning environment which do not manifest in simpler environments. We then provide a comparison of model-free, linear learning algorithms on this challenging problem set.

READ FULL TEXT
research
03/23/2022

An Optical Controlling Environment and Reinforcement Learning Benchmarks

Deep reinforcement learning has the potential to address various scienti...
research
06/30/2020

Evaluating the Performance of Reinforcement Learning Algorithms

Performance evaluations are critical for quantifying algorithmic advance...
research
11/25/2020

On limitations of learning algorithms in competitive environments

We discuss conceptual limitations of generic learning algorithms acting ...
research
01/07/2019

Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions

While the history of machine learning so far encompasses a series of pro...
research
04/26/2020

Reinforcement Learning Generalization with Surprise Minimization

Generalization remains a challenging problem for reinforcement learning ...
research
02/09/2023

Equivariant MuZero

Deep reinforcement learning repeatedly succeeds in closed, well-defined ...
research
07/30/2019

Model-Free Unsupervised Learning for Optimization Problems with Constraints

In many optimization problems in wireless communications, the expression...

Please sign up or login with your details

Forgot password? Click here to reset