SLACK: Stable Learning of Augmentations with Cold-start and KL regularization
CVPR 2023
Juliette Marrie1,2, Michael Arbel1, Diane Larlus1,2, Julien Mairal1
1 Univ.\ Grenoble Alpes, Inria, CNRS, Grenoble INP, LJK 2 NAVER LABS Europe
Visualization of learned policies
For different domains of the DomainNet dataset (one per line), we show an image from that domain (left) and that image transformed using the three most likely (middle) and the three least likely (right) augmentations for that domain, as estimated by SLACK
Bibtex
If you find our paper or pretrained models useful for your research, please consider citing us.
@InProceedings{Marrie_2023_CVPR, author = {Marrie, Juliette and Arbel, Michael and Larlus, Diane and Mairal, Julien}, title = {SLACK: Stable Learning of Augmentations with Cold-start and KL regularization}, booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, month = {June}, year = {2023} }