Preprint
Abstract
Real-world data streams naturally include the repetition of previous concepts. From a Continual Learning (CL) perspective, repetition is a property of the environment and, unlike replay, cannot be controlled by the agent. Nowadays, the Class-Incremental (CI) scenario represents the leading test-bed for assessing and comparing CL strategies. This scenario type is very easy to use, but it never allows revisiting previously seen classes, thus completely neglecting the role of repetition. We focus on the family of Class-Incremental with Repetition (CIR) scenario, where repetition is embedded in the definition of the stream. We propose two stochastic stream generators that produce a wide range of CIR streams starting from a single dataset and a few interpretable control parameters. We conduct the first comprehensive evaluation of repetition in CL by studying the behavior of existing CL strategies under different CIR streams. We then present a novel replay strategy that exploits repetition and counteracts the natural imbalance present in the stream. On both CIFAR100 and TinyImageNet, our strategy outperforms other replay approaches, which are not designed for environments with repetition.Keywords
All Related Versions
This publication has 7 references indexed in Scilit:
- Unified Probabilistic Deep Continual Learning through Generative Replay and Open Set RecognitionJournal of Imaging, 2022
- Imbalanced Continual Learning with Partitioning Reservoir Sampling2020
- Continual learning for robotics: Definition, framework, learning strategies, opportunities and challengesInformation Fusion, 2020
- Continuous learning in single-incremental-task scenariosNeural Networks, 2019
- Learning without ForgettingIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- Overcoming catastrophic forgetting in neural networksProceedings of the National Academy of Sciences of the United States of America, 2017
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998