Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
97 tokens/sec
GPT-4o
53 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Practical Transferability Estimation for Image Classification Tasks (2106.10479v3)

Published 19 Jun 2021 in cs.CV, cs.AI, and cs.LG

Abstract: Transferability estimation is an essential problem in transfer learning to predict how good the performance is when transferring a source model (or source task) to a target task. Recent analytical transferability metrics have been widely used for source model selection and multi-task learning. A major challenge is how to make transfereability estimation robust under the cross-domain cross-task settings. The recently proposed OTCE score solves this problem by considering both domain and task differences, with the help of transfer experiences on auxiliary tasks, which causes an efficiency overhead. In this work, we propose a practical transferability metric called JC-NCE score that dramatically improves the robustness of the task difference estimation in OTCE, thus removing the need for auxiliary tasks. Specifically, we build the joint correspondences between source and target data via solving an optimal transport problem with a ground cost considering both the sample distance and label distance, and then compute the transferability score as the negative conditional entropy of the matched labels. Extensive validations under the intra-dataset and inter-dataset transfer settings demonstrate that our JC-NCE score outperforms the auxiliary-task free version of OTCE for 7% and 12%, respectively, and is also more robust than other existing transferability metrics on average.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (37)
  1. Task2vec: Task embedding for meta-learning. pages 6430–6439, 2019.
  2. Geometric dataset distances via optimal transport. In Advances in Neural Information Processing Systems, volume 33, pages 21428–21439, 2020.
  3. An information-theoretic approach to transferability in task transfer learning. In 2019 IEEE International Conference on Image Processing (ICIP), pages 2309–2313. IEEE, 2019.
  4. Exploiting task relatedness for multiple task learning. In Learning Theory and Kernel Machines, pages 567–580. Springer, 2003.
  5. Analysis of representations for domain adaptation. Advances in neural information processing systems, 19:137–144, 2006.
  6. A theory of learning from different domains. Machine learning, 79(1-2):151–175, 2010.
  7. Learning bounds for domain adaptation. In Advances in neural information processing systems, pages 129–136, 2008.
  8. Large scale gan training for high fidelity natural image synthesis. International Conference on Learning Representations, 2019.
  9. Describing textures in the wild. In IEEE Conference on Computer Vision and Pattern Recognition, 2014.
  10. Decaf: A deep convolutional activation feature for generic visual recognition. In International conference on machine learning, pages 647–655, 2014.
  11. Unsupervised representation learning by predicting image rotations. In International Conference on Learning Representations, 2018.
  12. Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 580–587, 2014.
  13. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016.
  14. LV Kantorovich. On the translocation of masses, cr (dokl.) acad. Sci. URSS (NS), 37:199, 1942.
  15. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114, 2013.
  16. Alex Krizhevsky. Learning multiple layers of features from tiny images. Technical report, 2009.
  17. One-shot learning of object categories. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006.
  18. Domain adaptation: Learning bounds and algorithms. arXiv preprint arXiv:0902.3430, 2009.
  19. Andreas Maurer. Transfer bounds for linear feature learning. Machine learning, 75(3):327–350, 2009.
  20. Reading digits in natural images with unsupervised feature learning. In NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011, 2011.
  21. Leep: A new measure to evaluate transferability of learned representations. In International Conference on Machine Learning, 2020.
  22. M-E. Nilsback and A. Zisserman. Automated flower classification over a large number of classes. In Indian Conference on Computer Vision, Graphics and Image Processing, Dec 2008.
  23. Unsupervised learning of visual representations by solving jigsaw puzzles. In European Conference on Computer Vision, 2016.
  24. Cats and dogs. In IEEE Conference on Computer Vision and Pattern Recognition, 2012.
  25. Moment matching for multi-source domain adaptation. In Proceedings of the IEEE International Conference on Computer Vision, pages 1406–1415, 2019.
  26. Lorien Y Pratt. Discriminability-based transfer between neural networks. In Advances in neural information processing systems, pages 204–211, 1993.
  27. Adapting visual category models to new domains. In European conference on computer vision, pages 213–226. Springer, 2010.
  28. Meta-transfer learning for few-shot learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 403–412, 2019.
  29. Otce: A transferability metric for cross-domain cross-task representations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 15779–15788, June 2021.
  30. Transferability and hardness of supervised classification tasks. In Proceedings of the IEEE International Conference on Computer Vision, pages 1395–1405, 2019.
  31. Rotation equivariant cnns for digital pathology. In International Conference on Medical Image Computing and Computer-Assisted Intervention, 2018.
  32. Transfer learning via learning to transfer. In International Conference on Machine Learning, pages 5085–5094, 2018.
  33. How transferable are features in deep neural networks? In Advances in neural information processing systems, pages 3320–3328, 2014.
  34. Taskonomy: Disentangling task transfer learning. pages 3712–3722, 2018.
  35. S4l: Self-supervised semi-supervised learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 1476–1485, 2019a.
  36. A large-scale study of representation learning with the visual task adaptation benchmark. arXiv preprint arXiv:1910.04867, 2019b.
  37. Overcoming negative transfer: A survey. arXiv preprint arXiv:2009.00909, 2020.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Yang Tan (39 papers)
  2. Yang Li (1144 papers)
  3. Shao-Lun Huang (48 papers)
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now