Incremental Learning with Maximum Entropy Regularization: Rethinking Forgetting and Intransigence
This work addresses incremental learning challenges for AI systems that need to adapt to new data over time, representing an incremental improvement with novel regularization and sampling techniques.
The paper tackles forgetting and intransigence in incremental learning by proposing MEDIC, which uses Maximum Entropy Regularization and DropOut Sampling, and outperforms state-of-the-art methods in accuracy, forgetting, and intransigence on CIFAR100 and TinyImageNet datasets.
Incremental learning suffers from two challenging problems; forgetting of old knowledge and intransigence on learning new knowledge. Prediction by the model incrementally learned with a subset of the dataset are thus uncertain and the uncertainty accumulates through the tasks by knowledge transfer. To prevent overfitting to the uncertain knowledge, we propose to penalize confident fitting to the uncertain knowledge by the Maximum Entropy Regularizer (MER). Additionally, to reduce class imbalance and induce a self-paced curriculum on new classes, we exclude a few samples from the new classes in every mini-batch, which we call DropOut Sampling (DOS). We further rethink evaluation metrics for forgetting and intransigence in incremental learning by tracking each sample's confusion at the transition of a task since the existing metrics that compute the difference in accuracy are often misleading. We show that the proposed method, named 'MEDIC', outperforms the state-of-the-art incremental learning algorithms in accuracy, forgetting, and intransigence measured by both the existing and the proposed metrics by a large margin in extensive empirical validations on CIFAR100 and a popular subset of ImageNet dataset (TinyImageNet).