AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Preference Transformer: Modeling Human Preferences using Transformers for RL

Changyeon Kim, Jongjin Park, Jinwoo Shin, Honglak Lee, Pieter Abbeel, Kimin Lee

ICLR 2023 · Reinforcement learning › RL from feedback · approach: Reward modeling versus direct optimization · 4 citations (OpenAlex)

Read the paper

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
Use weighted sums of non-Markovian rewards with transformer preference attention instead of standard Markovian rewardsIQL reward learning from real human preferences · AntMazenormalized score: 20.16 -> 49.35Improved
Use weighted sums of non-Markovian rewards with transformer preference attention instead of standard Markovian rewardsIQL reward learning from real human preferences · Gym-Mujoco locomotionnormalized score: 62.43 -> 83.72Improved
Use weighted sums of non-Markovian rewards with transformer preference attention instead of standard Markovian rewardsIQL reward learning from real human preferences · Robosuitesuccess rate: 72.81 -> 74.66Mixed
Use transformer-based weighted non-Markovian preference modeling instead of LSTM non-Markovian reward modelingIQL reward learning from real human preferences · AntMazenormalized score: 24.29 -> 49.35Improved
Use transformer-based weighted non-Markovian preference modeling instead of LSTM non-Markovian reward modelingIQL reward learning from real human preferences · Gym-Mujoco locomotionnormalized score: 70.98 -> 83.72Improved
Use transformer-based weighted non-Markovian preference modeling instead of LSTM non-Markovian reward modelingIQL reward learning from real human preferences · Robosuitesuccess rate: 68.69 -> 74.66Improved
Train with real human rather than scripted preferencesPreference Transformer · AntMazenormalized score: 50.00 -> 49.35No 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