Diversity-Based Trajectory and Goal Selection with Hindsight Experience Replay

Tianhong Dai, Hengyan Liu, Kai Arulkumaran, Guangyu Ren, Anil Anthony Bharath

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

9 Citations (Scopus)

Abstract

Hindsight experience replay (HER) is a goal relabelling technique typically used with off-policy deep reinforcement learning algorithms to solve goal-oriented tasks; it is well suited to robotic manipulation tasks that deliver only sparse rewards. In HER, both trajectories and transitions are sampled uniformly for training. However, not all of the agent’s experiences contribute equally to training, and so naive uniform sampling may lead to inefficient learning. In this paper, we propose diversity-based trajectory and goal selection with HER (DTGSH). Firstly, trajectories are sampled according to the diversity of the goal states as modelled by determinantal point processes (DPPs). Secondly, transitions with diverse goal states are selected from the trajectories by using k-DPPs. We evaluate DTGSH on five challenging robotic manipulation tasks in simulated robot environments, where we show that our method can learn more quickly and reach higher performance than other state-of-the-art approaches on all tasks.

Original languageEnglish
Title of host publicationPRICAI 2021
Subtitle of host publicationTrends in Artificial Intelligence - 18th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2021, Proceedings
EditorsDuc Nghia Pham, Thanaruk Theeramunkong, Guido Governatori, Fenrong Liu
PublisherSpringer Science and Business Media Deutschland GmbH
Pages32-45
Number of pages14
ISBN (Print)9783030893699
DOIs
Publication statusPublished - 2021
Event18th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2021 - Virtual, Online
Duration: 8 Nov 202112 Nov 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13033 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference18th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2021
CityVirtual, Online
Period8/11/2112/11/21

Keywords

  • Deep reinforcement learning
  • Determinantal point processes
  • Hindsight experience replay

Fingerprint

Dive into the research topics of 'Diversity-Based Trajectory and Goal Selection with Hindsight Experience Replay'. Together they form a unique fingerprint.

Cite this