Abstract

Significance Deep neural networks are currently the most successful machine-learning technique for solving a variety of tasks, including language translation, image classification, and image generation. One weakness of such models is that, unlike humans, they are unable to learn multiple tasks sequentially. In this work we propose a practical solution to train such models sequentially by protecting the weights important for previous tasks. This approach, inspired by synaptic consolidation in neuroscience, enables state of the art results on multiple reinforcement learning problems experienced sequentially.

Keywords

ForgettingArtificial neural networkComputer sciencePsychologyArtificial intelligenceCognitive psychology

Affiliated Institutions

Related Publications

Neural GPUs Learn Algorithms

ukasz Kaiser , Ilya Sutskever

Abstract: Learning an algorithm from examples is a fundamental problem that has been widely studied. Recently it has been addressed using neural networks, in particular by Neura...

2016 arXiv (Cornell University) 63 citations

GeePS

Henggang Cui , Hao Zhang , Gregory R. Ganger +2 more

Large-scale deep learning requires huge computational resources to train a multi-layer neural network. Recent systems propose using 100s to 1000s of machines to train networks w...

2016 296 citations

Publication Info

Year
2017
Type
article
Volume
114
Issue
13
Pages
3521-3526
Citations
6331
Access
Closed

External Links

View on DOI.org

Social Impact

Altmetric
PlumX Metrics

Social media, news, blog, policy document mentions

Citation Metrics

6331
OpenAlex

Cite This

James Kirkpatrick, Razvan Pascanu, Neil C. Rabinowitz et al. (2017). Overcoming catastrophic forgetting in neural networks. Proceedings of the National Academy of Sciences , 114 (13) , 3521-3526. https://doi.org/10.1073/pnas.1611835114

Identifiers

DOI
10.1073/pnas.1611835114