Overcoming catastrophic forgetting in neural networks
Abstract
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.
- Publication:
-
Proceedings of the National Academy of Science
- Pub Date:
- March 2017
- DOI:
- 10.1073/pnas.1611835114
- arXiv:
- arXiv:1612.00796
- Bibcode:
- 2017PNAS..114.3521K
- Keywords:
-
- Computer Science - Machine Learning;
- Computer Science - Artificial Intelligence;
- Statistics - Machine Learning
- E-Print:
- doi:10.1073/pnas.1611835114