AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Adaptive Interest for Emphatic Reinforcement Learning

Martin Klissarov, Rasool Fakoor, Jonas Mueller, Kavosh Asadi, Taesup Kim, Alex Smola

NeurIPS 2022 · Reinforcement learning › Visual and representation RL · approach: In-context and adaptive RL · 0 citations (OpenAlex)

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
adapt the interest function with meta-gradients (MINT) instead of fixed interestemphatic TD / PPO · H4R8T off-policy evaluationMSVEImproved
use TD-error-based adaptive interest instead of meta-gradient interestemphatic TD · 4R8T and H4R8T off-policy evaluationMSVENo clear effect
add adaptive interest to ETDLBETDLB · 4R8T off-policy evaluationconvergence speedImproved
add MINT on top of PPOPPO · MinAtar (5 games)rewardImproved
use TD-error interest in controlPPO · MuJoCorewardNo clear effect
add MINT interest function to PPOPPO · MuJoCorewardImproved
add hypergradient descent learning-rate adaptationPPO · MuJoCorewardNo clear effect
transfer learned interest function to new taskactor-critic · FourRoomsTransferlearning speedImproved
transfer meta-learned intrinsic reward instead of interestactor-critic · FourRoomsTransferlearning speedNo clear effect

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