Publication

Dynamic Experience Replay

This paper presents a novel technique in reinforcement learning that largely improves the training efficiency, especially in contact-rich robotic assembly tasks.

Download publication

Abstract

Dynamic Experience Replay

Jieliang Luo, Hui Li

Conference on Robot Learning 2019

We present a novel technique called Dynamic Experience Replay (DER) that allows Reinforcement Learning (RL) algorithms to use experience replay samples not only from human demonstrations but also successful transitions generated by RL agents during training and therefore improve training efficiency. It can be combined with an arbitrary off-policy RL algorithm, such as DDPG or DQN, and their distributed versions. We build upon Ape-X DDPG and demonstrate our approach on robotic tight-fitting joint assembly tasks, based on force/torque and Cartesian pose observations. In particular, we run experiments on two different tasks: peg-in-hole and lap-joint. In each case, we compare different replay buffer structures and how DER affects them. Our ablation studies show that Dynamic Experience Replay is a crucial ingredient that either largely shortens the training time in these challenging environments or solves the tasks that the vanilla Ape-X DDPG cannot solve. We also show that our policies learned purely in simulation can be deployed successfully on the real robot.

Related Resources

Publication

2023

BOP-Elites: A Bayesian Optimisation Approach to Quality Diversity Search with Black-Box descriptor functions

An algorithm that efficiently tackles expensive black-box optimization…

Publication

2020

Learning to Simulate and Design for Structural Engineering

The structural design process for buildings is time consuming and…

Publication

2019

Relational Graph Representation Learning for Open-Domain Question Answering

We introduce a relational graph neural network with bi-directional…

Publication

2022

CAPRI-Net: Learning Compact CAD Shapes with Adaptive Primitive Assembly

We introduce CAPRI-Net, a self-supervised neural net-work for learning…

Get in touch

Something pique your interest? Get in touch if you’d like to learn more about Autodesk Research, our projects, people, and potential collaboration opportunities.

Contact us