Research

SpotTune: Transfer Learning through Adaptive Fine-tuning

Computer Vision

Authors

  • Yunhui Guo
  • Honghui Shi
  • Abhishek Kumar
  • Kristen Grauman
  • Tajana Rosing
  • Rogerio Feris

Published on

11/21/2018

Transfer learning, which allows a source task to affect the inductive bias of the target task, is widely used in computer vision. The typical way of conducting transfer learning with deep neural networks is to fine-tune a model pre-trained on the source task using data from the target task. In this paper, we propose an adaptive fine-tuning approach, called SpotTune, which finds the optimal fine-tuning strategy per instance for the target data. In SpotTune, given an image from the target task, a policy network is used to make routing decisions on whether to pass the image through the fine-tuned layers or the pre-trained layers. We conduct extensive experiments to demonstrate the effectiveness of the proposed approach. Our method outperforms the traditional fine-tuning approach on 12 out of 14 standard datasets.We also compare SpotTune with other state-of-the-art fine-tuning strategies, showing superior performance. On the Visual Decathlon datasets, our method achieves the highest score across the board without bells and whistles.

Please cite our work using the BibTeX below.

@article{DBLP:journals/corr/abs-1811-08737,
  author    = {Yunhui Guo and
               Honghui Shi and
               Abhishek Kumar and
               Kristen Grauman and
               Tajana Rosing and
               Rog{\'{e}}rio Schmidt Feris},
  title     = {SpotTune: Transfer Learning through Adaptive Fine-tuning},
  journal   = {CoRR},
  volume    = {abs/1811.08737},
  year      = {2018},
  url       = {http://arxiv.org/abs/1811.08737},
  archivePrefix = {arXiv},
  eprint    = {1811.08737},
  timestamp = {Wed, 16 Oct 2019 14:14:57 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-1811-08737.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

Close Modal