Deep Metric Learning to Rank
Abstract
We propose a novel deep metric learning method by revisiting the learning to rank approach. Our method, named FastAP, optimizes the rank-based Average Precision measure, using an approximation derived from distance quantization. FastAP has a low complexity compared to existing methods, and is tailored for stochastic gradient descent. To fully exploit the benefits of the ranking formulation, we also propose a new minibatch sampling scheme, as well as a simple heuristic to enable large-batch training. On three few-shot image retrieval datasets, FastAP consistently outperforms competing methods, which often involve complex optimization heuristics or costly model ensembles.
Cite
Text
Cakir et al. "Deep Metric Learning to Rank." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019. doi:10.1109/CVPR.2019.00196Markdown
[Cakir et al. "Deep Metric Learning to Rank." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019.](https://mlanthology.org/cvpr/2019/cakir2019cvpr-deep/) doi:10.1109/CVPR.2019.00196BibTeX
@inproceedings{cakir2019cvpr-deep,
title = {{Deep Metric Learning to Rank}},
author = {Cakir, Fatih and He, Kun and Xia, Xide and Kulis, Brian and Sclaroff, Stan},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {2019},
doi = {10.1109/CVPR.2019.00196},
url = {https://mlanthology.org/cvpr/2019/cakir2019cvpr-deep/}
}