In the dataset section, the common use-cases (CIFAR10H, Music, LableMe) are not well described, nor cited properly.
https://peerannot.github.io/datasets/
Once the labeling standardization has been performed, this should be a priority for you @AxelDubar .