MASIL: Towards Maximum Separable Class Representation for Few Shot Class Incremental Learning
This addresses the challenge of continual learning with limited annotated data, which is critical for real-world applications where data collection is costly, though it appears incremental in method.
The paper tackles the problem of forgetting old classes and overfitting to new ones in Few Shot Class Incremental Learning (FSCIL) by proposing MASIL, a framework that learns maximally separable classifier weights using a simplex Equiangular Tight Frame and concept factorization, achieving state-of-the-art results on benchmarks like miniImageNet, CIFAR-100, and CUB-200.
Few Shot Class Incremental Learning (FSCIL) with few examples per class for each incremental session is the realistic setting of continual learning since obtaining large number of annotated samples is not feasible and cost effective. We present the framework MASIL as a step towards learning the maximal separable classifier. It addresses the common problem i.e forgetting of old classes and over-fitting to novel classes by learning the classifier weights to be maximally separable between classes forming a simplex Equiangular Tight Frame. We propose the idea of concept factorization explaining the collapsed features for base session classes in terms of concept basis and use these to induce classifier simplex for few shot classes. We further adds fine tuning to reduce any error occurred during factorization and train the classifier jointly on base and novel classes without retaining any base class samples in memory. Experimental results on miniImageNet, CIFAR-100 and CUB-200 demonstrate that MASIL outperforms all the benchmarks.