the archive
#sparse rewards
3 resultscs.LG2019
Episodic Curiosity through Reachability
Nikolay Savinov, Anton Raichuk, Raphaël Marinier +4
The paper introduces an episodic curiosity method that measures novelty by the number of steps needed to reach a current observation from past observations, providing an intrinsic…
#curiosity-driven exploration#episodic memory#reachability#sparse rewards
cs.RO2019
Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards
Gerrit Schoettler, Ashvin Nair, Jianlan Luo +4
The paper presents a deep reinforcement learning approach that uses visual inputs and natural reward signals, such as sparse rewards and goal images, to learn controllers for indus…
#industrial insertion#visual reinforcement learning#sparse rewards#demonstration learning
cs.LG2019
On Hard Exploration for Reinforcement Learning: a Case Study in Pommerman
Chao Gao, Bilal Kartal, Pablo Hernandez-Leal +1
The paper investigates why model-free reinforcement learning struggles in the sparse-reward, multi-agent Pommerman environment and introduces a model-based reasoning module that pr…
#hard exploration#multi-agent reinforcement learning#pommerman#model-based exploration