Multi-task Self-Supervised Visual Learning
About
We investigate methods for combining multiple self-supervised tasks--i.e., supervised tasks where data can be collected without manual labeling--in order to train a single visual representation. First, we provide an apples-to-apples comparison of four different self-supervised tasks using the very deep ResNet-101 architecture. We then combine tasks to jointly train a network. We also explore lasso regularization to encourage the network to factorize the information in its representation, and methods for "harmonizing" network inputs in order to learn a more unified representation. We evaluate all methods on ImageNet classification, PASCAL VOC detection, and NYU depth prediction. Our results show that deeper networks work better, and that combining tasks--even via a naive multi-head architecture--always improves performance. Our best joint network nearly matches the PASCAL performance of a model pre-trained on ImageNet classification, and matches the ImageNet network on NYU depth prediction.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Object Detection | COCO 2017 (val) | AP32.7 | 2454 | |
| Image Classification | ImageNet-1k (val) | Top-1 Accuracy31.5 | 1453 | |
| Image Classification | ImageNet (val) | -- | 1206 | |
| Object Detection | PASCAL VOC 2007 (test) | mAP70.5 | 821 | |
| Depth Estimation | NYU v2 (test) | Threshold Accuracy (delta < 1.25)79.3 | 423 | |
| Image Classification | ImageNet (val) | -- | 354 | |
| Image Classification | ImageNet | -- | 55 | |
| Image Classification | VTAB v2 (test) | Mean Accuracy59.2 | 39 | |
| Depth Prediction | NYU Depth | -- | 5 |