AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

On the Importance of Exploration for Generalization in Reinforcement Learning

Yiding Jiang, J Zico Kolter, Roberta Raileanu

NeurIPS 2023 · Reinforcement learning › Visual and representation RL · approach: Learned state and action representations

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add TEE (training on diverse data) to QR-DQN without UCB (ablation)QR-DQN · Procgennormalized scoreImproved
add epistemic-uncertainty UCB exploration to QR-DQN (ablation)QR-DQN · Procgennormalized scoreImproved
add TEE on top of QR-DQN+UCB (full method) (ablation)QR-DQN+UCB · Procgennormalized scoreImproved
explore with total uncertainty instead of epistemic uncertainty (ablation)QR-DQN+UCB · Procgentest scoreHurt
use Bootstrapped DQN or DQN+UCB (total uncertainty) instead of plain DQNDQN · Procgennormalized scoreHurt
use epsilon-z-greedy exploration on QR-DQNQR-DQN · Procgennormalized scoreNo clear effect
use UA-DQN instead of deep ensemble UCBQR-DQN+UCB · Procgennormalized scoreHurt
use NoisyNet exploration instead of EDEQR-DQN · Procgennormalized scoreHurt
use EDE instead of QR-DQNProcgen agent · Procgenmedian score: almost 3 times the median scoreImproved
use EDE instead of QR-DQNProcgen agent · ProcgenIQM: more than 2 times the IQMImproved
use EDE instead of Rainbow explorationRainbow · Crafterscore (%): 4.3 -> 11.7Improved
use EDE instead of DreamerV2Crafter agent · Crafterscore (%): 10.0 -> 11.7Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy