Learning Bayesian Sparse Networks with Full Experience Replay for Continual Learning
Authors:
- 1 June 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 109-118
Abstract
Continual Learning (CL) methods aim to enable machine learning models to learn new tasks without catastrophic forgetting of those that have been previously mastered. Existing CL approaches often keep a buffer of previously-seen samples, perform knowledge distillation, or use regularization techniques towards this goal. Despite their performance, they still suffer from interference across tasks which leads to catastrophic forgetting. To ameliorate this problem, we propose to only activate and select sparse neurons for learning current and past tasks at any stage. More parameters space and model capacity can thus be reserved for the future tasks. This minimizes the interference between parameters for different tasks. To do so, we propose a Sparse neural Network for Continual Learning (SNCL), which employs variational Bayesian sparsity priors on the activations of the neurons in all layers. Full Experience Replay (FER) provides effective supervision in learning the sparse activations of the neurons in different layers. A loss-aware reservoir-sampling strategy is developed to maintain the memory buffer. The proposed method is agnostic as to the network structures and the task boundaries. Experiments on different datasets show that SNCL achieves state-of-the-art result for mitigating forgetting.Keywords
All Related Versions
This publication has 16 references indexed in Scilit:
- iTAML: An Incremental Task-Agnostic Meta-learning Approach2020
- Learning a Unified Classifier Incrementally via Rebalancing2019
- Variational Bayesian Dropout With a Hierarchical Prior2019
- Learning without ForgettingIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- iCaRL: Incremental Classifier and Representation Learning2017
- Overcoming catastrophic forgetting in neural networksProceedings of the National Academy of Sciences of the United States of America, 2017
- Deep Residual Learning for Image Recognition2016
- Child: A First Step Towards Continual Learning1998
- Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem1989
- Random sampling with a reservoirACM Transactions on Mathematical Software, 1985