• Classified by Topic • Classified by Publication Type • Sorted by Date • Sorted by First Author Last Name • Classified by Funding Source •
t-DGR: A Trajectory-Based Deep Generative Replay Method for Continual Learning in Decision Making.
William Yue,
Bo Liu, and Peter Stone.
In
Conference on Lifelong Learning Agents (CoLLAs), July 2024.
[PDF]599.7kB [poster.pdf]709.1kB
Deep generative replay has emerged as a promising approach for continual learningin decision-making tasks. This approach addresses the problem of catastrophicforgetting by leveraging the generation of trajectories from previouslyencountered tasks to augment the current dataset. However, existing deepgenerative replay methods for continual learning rely on autoregressive models,which suffer from compounding errors in the generated trajectories. In thispaper, we propose a simple, scalable, and non-autoregressive method for continuallearning in decision-making tasks using a generative model that generates tasksamples conditioned on the trajectory timestep. We evaluate our method onContinual World benchmarks and find that our approach achieves state-of-the-artperformance on the average success rate metric among continual learning methods.
@InProceedings{yue2024tdgr,
author = {William Yue and Bo Liu and Peter Stone},
title = {t-DGR: A Trajectory-Based Deep Generative Replay Method for Continual Learning in Decision Making},
booktitle = {Conference on Lifelong Learning Agents (CoLLAs)},
year = {2024},
month = {July},
location = {Pisa, Italy},
abstract = {Deep generative replay has emerged as a promising approach for continual learning
in decision-making tasks. This approach addresses the problem of catastrophic
forgetting by leveraging the generation of trajectories from previously
encountered tasks to augment the current dataset. However, existing deep
generative replay methods for continual learning rely on autoregressive models,
which suffer from compounding errors in the generated trajectories. In this
paper, we propose a simple, scalable, and non-autoregressive method for continual
learning in decision-making tasks using a generative model that generates task
samples conditioned on the trajectory timestep. We evaluate our method on
Continual World benchmarks and find that our approach achieves state-of-the-art
performance on the average success rate metric among continual learning methods.
},
}
Generated by bib2html.pl (written by Patrick Riley ) on Sun Nov 24, 2024 20:24:53