iCaRL: Incremental Classifier and Representation Learning
This addresses the challenge of building AI systems that can learn continuously over time, which is crucial for real-world applications, though it builds on prior incremental learning work.
The paper tackles the problem of incremental learning in AI by introducing iCaRL, a strategy that enables learning new classes progressively from a data stream, and it demonstrates that iCaRL outperforms other methods on datasets like CIFAR-100 and ImageNet ILSVRC 2012, achieving long-term learning where others fail.
A major open problem on the road to artificial intelligence is the development of incrementally learning systems that learn about more and more concepts over time from a stream of data. In this work, we introduce a new training strategy, iCaRL, that allows learning in such a class-incremental way: only the training data for a small number of classes has to be present at the same time and new classes can be added progressively. iCaRL learns strong classifiers and a data representation simultaneously. This distinguishes it from earlier works that were fundamentally limited to fixed data representations and therefore incompatible with deep learning architectures. We show by experiments on CIFAR-100 and ImageNet ILSVRC 2012 data that iCaRL can learn many classes incrementally over a long period of time where other strategies quickly fail.