Year: 2021
Type: conference-paper
Abstract: How Well Does It Work?I In the OpenAI Gym CartPole-v0 benchmark vs. PPO, DQN I NE converges faster, has lower variance, lower regret I NE is more efficient, reliable, and safer 12 I In a double-pole benchmark vs. Sarsa, Q-MLP, etc.I The only method that can find solutions to 1m, 0.1m, POMDP 17 8/60
Author: Risto Miikkulainen
Institution: The University of Texas at Austin
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4205203293