FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learning
This addresses the challenge of incremental learning without storing exemplars, which is crucial for applications with privacy or memory constraints, though it appears incremental in approach.
The paper tackles the problem of catastrophic forgetting in exemplar-free class-incremental learning by introducing FeTrIL, which uses a fixed feature extractor and a pseudo-features generator to balance stability and plasticity, achieving faster incremental processes and outperforming ten existing methods in most cases on three datasets.
Exemplar-free class-incremental learning is very challenging due to the negative effect of catastrophic forgetting. A balance between stability and plasticity of the incremental process is needed in order to obtain good accuracy for past as well as new classes. Existing exemplar-free class-incremental methods focus either on successive fine tuning of the model, thus favoring plasticity, or on using a feature extractor fixed after the initial incremental state, thus favoring stability. We introduce a method which combines a fixed feature extractor and a pseudo-features generator to improve the stability-plasticity balance. The generator uses a simple yet effective geometric translation of new class features to create representations of past classes, made of pseudo-features. The translation of features only requires the storage of the centroid representations of past classes to produce their pseudo-features. Actual features of new classes and pseudo-features of past classes are fed into a linear classifier which is trained incrementally to discriminate between all classes. The incremental process is much faster with the proposed method compared to mainstream ones which update the entire deep model. Experiments are performed with three challenging datasets, and different incremental settings. A comparison with ten existing methods shows that our method outperforms the others in most cases.