In this paper, we propose a new method to overcome catastrophic forgetting by adding generative regularization to Bayesian inference framework. Bayesian method provides a general framework for continual learning. We could further construct a generative regularization term for all given classification models by leveraging energy-based models and Langevin-dynamic sampling to enrich the features learned in each task. By combining discriminative and generative loss together, we empirically show that the proposed method outperforms state-of-the-art methods on a variety of tasks, avoiding catastrophic forgetting in continual learning. In particular, the proposed method outperforms baseline methods over 15% on the Fashion-MNIST dataset and 10% on the CUB dataset
@article{arxiv.1912.01238,
title = {Overcoming Catastrophic Forgetting by Generative Regularization},
author = {Patrick H. Chen and Wei Wei and Cho-jui Hsieh and Bo Dai},
journal= {arXiv preprint arXiv:1912.01238},
year = {2021}
}