- A network with cross-stitch units can learn an optimal combination of shared and task-specific representations.
- We propose a new sharing unit: “cross-stitch” unit.
- The units combine the activations from multiple networks and can be trained end-to-end.
- In the paper, we propose a principled approach to learn shared representations in ConvNets using multi-task learning.
- Abstract: Multi-task learning in Convolutional Networks has displayed remarkable success in the field of recognition.
Read the full article, click here.
@quantombone: “Cross-stitch networks for multi-task learning #deeplearning #computervision #CVPR2016”
One hundred percent of your contribution will fund improvements and new initiatives to benefit arXiv’s global scientific community. Please join the Simons Foundation and our generous member organizations and research labs in supporting arXiv.
[1604.03539] Cross-stitch Networks for Multi-task Learning