Cross-stitch Networks for Multi-task Learning - Robotics Institute Carnegie Mellon University

Cross-stitch Networks for Multi-task Learning

Ishan Misra, Abhinav Shrivastava, Abhinav Gupta, and Martial Hebert
Conference Paper, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, June, 2016

Abstract

Multi-task learning in Convolutional Networks has displayed remarkable success in the field of recognition. This success can be largely attributed to learning shared representations from multiple supervisory tasks. However, existing multi-task approaches rely on enumerating multiple network architectures specific to the tasks at hand, that do not generalize. In this paper, we propose a principled approach to learn shared representations in ConvNets using multi-task learning. Specifically, we propose a new sharing unit: "cross-stitch" unit. These units combine the activations from multiple networks and can be trained end-to-end. A network with cross-stitch units can learn an optimal combination of shared and task-specific representations. Our proposed method generalizes across multiple tasks and shows dramatically improved performance over baseline methods for categories with few training examples

BibTeX

@conference{Misra-2016-4855,
author = {Ishan Misra and Abhinav Shrivastava and Abhinav Gupta and Martial Hebert},
title = {Cross-stitch Networks for Multi-task Learning},
booktitle = {Proceedings of IEEE Conference on Computer Vision and Pattern Recognition},
year = {2016},
month = {June},
editor = {ONR MURI N000141612007; US Army Research Lab (ARL) CTA program (Agreement W911NF-10-2-0016)},
publisher = {IEEE},
}