 |
ML Newsnew | past | best | rss | submit | about |
| Overcoming catastrophic forgetting in neural networks (arxiv.org) | |
| 2 points by RushAndAPush on Dec 5, 2016 | hide | past | pdf | discuss on HN | |
|
In plain words: When a network learns a new task, it slows changes to the connections that mattered for earlier tasks, so old skills stay intact. Trained one after another on digit sorting and Atari games, it kept old abilities instead of forgetting them. AbstractThe ability to learn tasks in a sequential fashion is crucial to the development of artificial intelligence. Neural networks are not, in general, capable of this and it has been widely thought that catastrophic forgetting is an inevitable feature of connectionist models. We show that it is possible to overcome this limitation and train networks that can maintain expertise on tasks which they have not experienced for a long time. Our approach remembers old tasks by selectively slowing down learning on the weights important for those tasks. We demonstrate our approach is scalable and effective by solving a set of classification tasks based on the MNIST hand written digit dataset and by learning several Atari 2600 games sequentially. | |
| add comment on HN |