Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Reward Uncertainty for Exploration in Preference-based Reinforcement Learning

About

Conveying complex objectives to reinforcement learning (RL) agents often requires meticulous reward engineering. Preference-based RL methods are able to learn a more flexible reward model based on human preferences by actively incorporating human feedback, i.e. teacher's preferences between two clips of behaviors. However, poor feedback-efficiency still remains a problem in current preference-based RL algorithms, as tailored human feedback is very expensive. To handle this issue, previous methods have mainly focused on improving query selection and policy initialization. At the same time, recent exploration methods have proven to be a recipe for improving sample-efficiency in RL. We present an exploration method specifically for preference-based RL algorithms. Our main idea is to design an intrinsic reward by measuring the novelty based on learned reward. Specifically, we utilize disagreement across ensemble of learned reward models. Our intuition is that disagreement in learned reward model reflects uncertainty in tailored human feedback and could be useful for exploration. Our experiments show that exploration bonus from uncertainty in learned reward improves both feedback- and sample-efficiency of preference-based RL algorithms on complex robot manipulation tasks from MetaWorld benchmarks, compared with other existing exploration methods that measure the novelty of state visitation.

Xinran Liang, Katherine Shu, Kimin Lee, Pieter Abbeel• 2022

Related benchmarks

TaskDatasetResultRank
door-openMeta-World
Door Open Success Rate100
20
window-openMeta-World window-open
ASR70
20
window-closeMeta-World window-close
ASR80
20
door-unlockMeta-World
Success Rate48
14
Handle PressMeta-World
Success Rate58
14
door-lockMeta-World
Success Rate54
14
Door CloseMeta-World sim
Success Rate100
6
Box OpenUR5 real
Success Rate40
6
door-openUR5 real
Success Rate40
6
box-closeMeta-World sim
Box Close Success Rate60
6
Showing 10 of 13 rows

Other info

Follow for update