Command Palette
Search for a command to run...
Sylvestre-Alvise Rebuffi; Alexander Kolesnikov; Georg Sperl; Christoph H. Lampert
Abstract
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.
Code Repositories
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| class-incremental-learning-on-cifar100 | iCaRL | 10-stage average accuracy: 63.24 |
| incremental-learning-on-cifar-100-50-classes-2 | iCaRL* | Average Incremental Accuracy: 52.57 |
| incremental-learning-on-cifar-100-50-classes-3 | iCaRL* | Average Incremental Accuracy: 57.17 |
| incremental-learning-on-cifar-100-50-classes-4 | iCaRL | Average Incremental Accuracy: 71.33 |
| incremental-learning-on-cifar-100-b0-5steps | iCaRL | Average Incremental Accuracy: 71.14 |
| incremental-learning-on-imagenet-10-steps | iCaRL | # M Params: 11.68 Average Incremental Accuracy: 38.40 Average Incremental Accuracy Top-5: 63.70 Final Accuracy: 22.70 Final Accuracy Top-5: 44.00 |
| incremental-learning-on-imagenet-100-50-3 | iCaRL* | Average Incremental Accuracy: 65.56 |
| incremental-learning-on-imagenet100-10-steps | iCaRL | # M Params: 11.22 Average Incremental Accuracy Top-5: 83.60 Final Accuracy Top-5: 63.80 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.