Metalearning Continual Learning Algorithms

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Irie, Kazuki, Csordás, Róbert, Schmidhuber, Jürgen
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913693519839232
author Irie, Kazuki
Csordás, Róbert
Schmidhuber, Jürgen
author_facet Irie, Kazuki
Csordás, Róbert
Schmidhuber, Jürgen
contents General-purpose learning systems should improve themselves in open-ended fashion in ever-changing environments. Conventional learning algorithms for neural networks, however, suffer from catastrophic forgetting (CF), i.e., previously acquired skills are forgotten when a new task is learned. Instead of hand-crafting new algorithms for avoiding CF, we propose Automated Continual Learning (ACL) to train self-referential neural networks to metalearn their own in-context continual (meta)learning algorithms. ACL encodes continual learning (CL) desiderata -- good performance on both old and new tasks -- into its metalearning objectives. Our experiments demonstrate that ACL effectively resolves "in-context catastrophic forgetting," a problem that naive in-context learning algorithms suffer from; ACL-learned algorithms outperform both hand-crafted learning algorithms and popular meta-continual learning methods on the Split-MNIST benchmark in the replay-free setting, and enables continual learning of diverse tasks consisting of multiple standard image classification datasets. We also discuss the current limitations of in-context CL by comparing ACL with state-of-the-art CL methods that leverage pre-trained models. Overall, we bring several novel perspectives into the long-standing problem of CL.
format Preprint
id arxiv_https___arxiv_org_abs_2312_00276
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Metalearning Continual Learning Algorithms
Irie, Kazuki
Csordás, Róbert
Schmidhuber, Jürgen
Machine Learning
General-purpose learning systems should improve themselves in open-ended fashion in ever-changing environments. Conventional learning algorithms for neural networks, however, suffer from catastrophic forgetting (CF), i.e., previously acquired skills are forgotten when a new task is learned. Instead of hand-crafting new algorithms for avoiding CF, we propose Automated Continual Learning (ACL) to train self-referential neural networks to metalearn their own in-context continual (meta)learning algorithms. ACL encodes continual learning (CL) desiderata -- good performance on both old and new tasks -- into its metalearning objectives. Our experiments demonstrate that ACL effectively resolves "in-context catastrophic forgetting," a problem that naive in-context learning algorithms suffer from; ACL-learned algorithms outperform both hand-crafted learning algorithms and popular meta-continual learning methods on the Split-MNIST benchmark in the replay-free setting, and enables continual learning of diverse tasks consisting of multiple standard image classification datasets. We also discuss the current limitations of in-context CL by comparing ACL with state-of-the-art CL methods that leverage pre-trained models. Overall, we bring several novel perspectives into the long-standing problem of CL.
title Metalearning Continual Learning Algorithms
topic Machine Learning
url https://arxiv.org/abs/2312.00276