AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Outcome-Driven Reinforcement Learning via Variational Inference

Tim G. J. Rudner, Vitchyr H. Pong, Rowan Thomas McAllister, Yarin Gal, Sergey Levine

NeurIPS 2021 · Reinforcement learning › Visual and representation RL · approach: Hierarchical skills and priors · 3 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
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · Fetchfinal normalized distance (x100, lower is better): 19 -> 53Hurt
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · Faucetfinal normalized distance (x100, lower is better): 13 -> 37Hurt
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · Windowfinal normalized distance (x100, lower is better): 5.4 -> 7.9Hurt
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · Antfinal normalized distance (x100, lower is better): 9 -> 12Hurt
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · 2Dfinal normalized distance (x100, lower is better): 1.7 -> 1.0No clear effect
Fix q_T to the prior (remove dynamic discount factor) (ablation)ODAC · Pushfinal normalized distance (x100, lower is better): 35 -> 37No clear effect
Replace learned dynamics model with fixed-variance Laplace model (ablation)ODAC · Fetchfinal normalized distance (x100, lower is better): 19 -> 15No clear effect
Replace learned dynamics model with fixed-variance Laplace model (ablation)ODAC · Antfinal normalized distance (x100, lower is better): 9 -> 11No clear effect
Replace learned dynamics model with fixed-variance Laplace model (ablation)ODAC · Faucetfinal normalized distance (x100, lower is better): 13 -> 15No clear effect
Fix both q_T and dynamics model (ablation)ODAC · Fetchfinal normalized distance (x100, lower is better): 19 -> 66Hurt
Fix both q_T and dynamics model (ablation)ODAC · Faucetfinal normalized distance (x100, lower is better): 13 -> 38Hurt
Use learned dynamics model for both rewards and model-based planning (ablation)ODAC · six environmentsfinal distanceHurt

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