TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

SotA/Methodology/Continual Learning

Continual Learning

75 benchmarks2644 papers

Continual Learning (also known as Incremental Learning, Life-long Learning) is a concept to learn a model for a large number of tasks sequentially without forgetting knowledge obtained from the preceding tasks, where the data in the old tasks are not available anymore during training new ones.
If not mentioned, the benchmarks here are Task-CL, where task-id is provided on validation.

Source:
Continual Learning by Asymmetric Loss Approximation with Single-Side Overestimation
Three scenarios for continual learning
Lifelong Machine Learning
Continual lifelong learning with neural networks: A review

Benchmarks

Continual Learning on PASCAL VOC 2012

mIoUMean IoU (val)Mean IoUMean IoU (test)

Continual Learning on ADE20K

mIoUMean IoU (test)

Continual Learning on ASC (19 tasks)

F1 - macro

Continual Learning on visual domain decathlon (10 tasks)

decathlon discipline (Score)Avg. Accuracy

Continual Learning on mini-Imagenet

Last Accuracy Average Accuracy

Continual Learning on CIFAR-100

Last AccuracyAverage Accuracy

Continual Learning on Cifar100 (20 tasks)

Average Accuracy

Continual Learning on Tiny-ImageNet (10tasks)

Average Accuracy

Continual Learning on F-CelebA (10 tasks)

Acc

Continual Learning on cifar100

10-stage average accuracyAverage accuracy - 5 tasksaverage accuracy - 10 tasksaverage accuracy - 20 tasks

Continual Learning on CUB-200-2011

Last Accuracy Average Accuracy

Continual Learning on 20Newsgroup (10 tasks)

F1 - macro

Continual Learning on CUBS (Fine-grained 6 Tasks)

Accuracy

Continual Learning on DSC (10 tasks)

F1 - macro

Continual Learning on Flowers (Fine-grained 6 Tasks)

Accuracy

Continual Learning on ImageNet (Fine-grained 6 Tasks)

Accuracy

Continual Learning on Sketch (Fine-grained 6 Tasks)

Accuracy

Continual Learning on Stanford Cars (Fine-grained 6 Tasks)

Accuracy

Continual Learning on Wikiart (Fine-grained 6 Tasks)

Accuracy

Continual Learning on Cifar100 (10 tasks)

Average Accuracy

Continual Learning on ImageNet-50 (5 tasks)

Accuracy

Continual Learning on TiROD

Omega

Continual Learning on Permuted MNIST

Average AccuracyBWTMLP Hidden Layers-widthPretrained/Transfer Learning

Continual Learning on CIFAR-100 - 50 classes + 10 steps of 5 classes

Final Accuracy

Continual Learning on CIFAR-100 - 50 classes + 5 steps of 10 classes

Final Accuracy

Continual Learning on Cityscapes

mIoU

Continual Learning on split CIFAR-100

Average AccuracyBWT

Continual Learning on 2010 i2b2/VA

F1 (macro)F1 (micro)

Continual Learning on 5-Datasets

Average AccuracyBWT

Continual Learning on 5-dataset - 1 epoch

Accuracy

Continual Learning on AIDS

1:3 Accuracy

Continual Learning on CIFAR-100 AlexNet - 300 Epoch

Accuracy

Continual Learning on CIFAR-100 ResNet-18 - 300 Epochs

Accuracy

Continual Learning on CIFAR100-B0(50 tasks)-no-exemplars

Average Incremental Accuracy

Continual Learning on CUB-200-2011 (20 tasks) - 1 epoch

Accuracy

Continual Learning on Cifar100 (20 tasks) - 1 epoch

Average Accuracy

Continual Learning on Cifar100-B0(10 tasks)-no-exemplars

Average Incremental Accuracy

Continual Learning on Cifar100-B0(20 tasks)-no-exemplars

Average Incremental Accuracy

Continual Learning on Coarse-CIFAR100

Average Accuracy

Continual Learning on ImageNetSubset

Average accuracy - 5 tasksaverage accuracy - 10 tasksaverage accuracy - 20 tasks

Continual Learning on MLT17

Acc

Continual Learning on MiniImageNet ResNet-18 - 300 Epochs

Accuracy

Continual Learning on OntoNotes 5.0

F1 (macro)F1 (micro)

Continual Learning on Rotated MNIST

Average Accuracy

Continual Learning on Split CIFAR-10 (5 tasks)

Top 1 Accuracy %

Continual Learning on Split MNIST (5 tasks)

Top 1 Accuracy %

Continual Learning on TinyImageNet

Average accuracy - 5 tasksaverage accuracy - 10 tasksaverage accuracy - 20 tasks

Continual Learning on TinyImageNet ResNet-18 - 300 Epochs

Accuracy

Continual Learning on conll2003

F1 (macro)F1 (micro)

Continual Learning on mini-Imagenet (20 tasks) - 1 epoch

Accuracy

Continual Learning on miniImagenet

Average AccuracyBWT