PT Unknown
AU Alex Gomez-Villa
   Bartlomiej Twardowski
   Lu Yu
   Andrew Bagdanov
   Joost Van de Weijer
TI Continually Learning Self-Supervised Representations With Projected Functional Regularization
BT CVPR 2022 Workshop on Continual Learning (CLVision, 3rd Edition)
PY 2022
BP 3866
EP 3876
DI 10.1109/CVPRW56347.2022.00432
DE Computer vision; Conferences; Self-supervised learning; Image representation; Pattern recognition
AB Recent self-supervised learning methods are able to learn high-quality image representations and are closing the gap with supervised approaches. However, these methods are unable to acquire new knowledge incrementally – they are, in fact, mostly used only as a pre-training phase over IID data. In this work we investigate self-supervised methods in continual learning regimes without any replaymechanism. We show that naive functional regularization,also known as feature distillation, leads to lower plasticity and limits continual learning performance. Instead, we propose Projected Functional Regularization in which a separate temporal projection network ensures that the newly learned feature space preserves information of the previous one, while at the same time allowing for the learning of new features. This prevents forgetting while maintaining the plasticity of the learner. Comparison with other incremental learning approaches applied to self-supervision demonstrates that our method obtains competitive performance indifferent scenarios and on multiple datasets.
ER