TL;DR
This paper introduces Smooth-AP, a differentiable approximation of Average Precision that enables end-to-end training of deep networks for large-scale image retrieval, outperforming existing methods on multiple benchmarks.
Contribution
The paper proposes Smooth-AP, a novel differentiable objective for optimizing AP directly, facilitating scalable and effective deep metric learning for image retrieval.
Findings
Smooth-AP improves retrieval performance on multiple benchmarks.
It is especially effective on large-scale datasets.
Smooth-AP enables end-to-end training with simple implementation.
Abstract
Optimising a ranking-based metric, such as Average Precision (AP), is notoriously challenging due to the fact that it is non-differentiable, and hence cannot be optimised directly using gradient-descent methods. To this end, we introduce an objective that optimises instead a smoothed approximation of AP, coined Smooth-AP. Smooth-AP is a plug-and-play objective function that allows for end-to-end training of deep networks with a simple and elegant implementation. We also present an analysis for why directly optimising the ranking based metric of AP offers benefits over other deep metric learning losses. We apply Smooth-AP to standard retrieval benchmarks: Stanford Online products and VehicleID, and also evaluate on larger-scale datasets: INaturalist for fine-grained category retrieval, and VGGFace2 and IJB-C for face retrieval. In all cases, we improve the performance over the…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
