Overcoming catastrophic forgetting with hard attention to the task

Catastrophic forgetting occurs when a neural network loses the information learned in a previous task after training on subsequent tasks. This problem remains a hurdle for artificial intelligence systems with sequential learning capabilities. In this paper, we propose a task-based hard attention mechanism that preserves previous tasks' information without affecting the current task's learning. A hard attention mask is learned concurrently to every task, through stochastic gradient descent, and previous masks are exploited to condition such learning. We show that the proposed mechanism is effective for reducing catastrophic forgetting, cutting current rates by 45 to 80%. We also show that it is robust to different hyperparameter choices, and that it offers a number of monitoring capabilities. The approach features the possibility to control both the stability and compactness of the learned knowledge, which we believe makes it also attractive for online learning or network compression applications.

PDF Abstract ICML 2018 PDF ICML 2018 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Continual Learning 20Newsgroup (10 tasks) HAT F1 - macro 0.9521 # 2
Continual Learning ASC (19 tasks) HAT F1 - macro 0.7816 # 7
Continual Learning DSC (10 tasks) HAT F1 - macro 0.8614 # 3
Continual Learning F-CelebA (10 tasks) HAT Acc 0.5673 # 6


No methods listed for this paper. Add relevant methods here