atlas

Research map / Reinforcement learning / Visual and representation RL

Automatic Curriculum Learning through Value Disagreement

NeurIPS 2020 · Reinforcement learning › Visual and representation RL · approach: Curriculum and environment design · 39 citations (OpenAlex)

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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
sample training goals by value-ensemble disagreement (VDS) instead of uniform samplingHER with DDPG · 13 Fetch/Hand robotics tasks and 3 mazessuccess rate / sample efficiencyHelped
use VDS instead of GoalGAN goal proposalgoal-conditioned RL · Ant navigation (2 environments)success rate / sample efficiencyHelped
change sampling function f (exp, tanh, square) instead of identity (ablation)VDS · Maze environmentssuccess rateNo clear effect
increase value ensemble size from 3 to 10 (ablation)VDS · Maze navigationsuccess rateNo clear effect
add VDS to vanilla DDPG without HER (ablation)DDPG · two environmentssuccess rateHelped
combine VDS with HER versus HER alone (ablation)HER · majority of environmentssample efficiencyHelped

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