Learning Efficient Models From Few Labels By Distillation From Multiple Tasks

Kenneth Borup, Cheng Perng Phoo, Bharath Hariharan

Keywords: transfer learning, semi-supervised learning, multi-source distillation

Abstract: We address the challenge of getting efficient yet accurate recognition systems that can be trained with limited labels. Many specialized applications of computer vision (e.g. analyzing X-rays or satellite images) have severe resource constraints both during training and inference. While transfer learning is an effective solution for training on small labeled datasets it still often requires a large base model for fine-tuning. In this paper we present a weighted multi-source distillation method; we distill multiple (diverse) source models trained on different domains, weighted by their relevance for the target task, into a single efficient model using limited labeled data. When the goal is accurate recognition under computational constraints, our approach outperforms both transfer learning from strong ImageNet initializations as well as state-of-the-art semi-supervised techniques such as FixMatch. When averaged over 8 diverse target tasks our method outperform the baselines by 5.6%-points and 4.5%-points, respectively.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning

TL;DR: We create an efficient model for a novel task via task similarity-weighted multi-source distillation.