Saved in:
Bibliographic Details
Main Authors: Rolínek, Michal, Musil, Vít, Paulus, Anselm, Vlastelica, Marin, Michaelis, Claudio, Martius, Georg
Format: Preprint
Published: 2019
Subjects:
Online Access:https://arxiv.org/abs/1912.03500
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909425766236160
author Rolínek, Michal
Musil, Vít
Paulus, Anselm
Vlastelica, Marin
Michaelis, Claudio
Martius, Georg
author_facet Rolínek, Michal
Musil, Vít
Paulus, Anselm
Vlastelica, Marin
Michaelis, Claudio
Martius, Georg
contents Rank-based metrics are some of the most widely used criteria for performance evaluation of computer vision models. Despite years of effort, direct optimization for these metrics remains a challenge due to their non-differentiable and non-decomposable nature. We present an efficient, theoretically sound, and general method for differentiating rank-based metrics with mini-batch gradient descent. In addition, we address optimization instability and sparsity of the supervision signal that both arise from using rank-based metrics as optimization targets. Resulting losses based on recall and Average Precision are applied to image retrieval and object detection tasks. We obtain performance that is competitive with state-of-the-art on standard image retrieval datasets and consistently improve performance of near state-of-the-art object detectors. The code is available at https://github.com/martius-lab/blackbox-backprop
format Preprint
id arxiv_https___arxiv_org_abs_1912_03500
institution arXiv
publishDate 2019
record_format arxiv
spellingShingle Optimizing Rank-based Metrics with Blackbox Differentiation
Rolínek, Michal
Musil, Vít
Paulus, Anselm
Vlastelica, Marin
Michaelis, Claudio
Martius, Georg
Machine Learning
Rank-based metrics are some of the most widely used criteria for performance evaluation of computer vision models. Despite years of effort, direct optimization for these metrics remains a challenge due to their non-differentiable and non-decomposable nature. We present an efficient, theoretically sound, and general method for differentiating rank-based metrics with mini-batch gradient descent. In addition, we address optimization instability and sparsity of the supervision signal that both arise from using rank-based metrics as optimization targets. Resulting losses based on recall and Average Precision are applied to image retrieval and object detection tasks. We obtain performance that is competitive with state-of-the-art on standard image retrieval datasets and consistently improve performance of near state-of-the-art object detectors. The code is available at https://github.com/martius-lab/blackbox-backprop
title Optimizing Rank-based Metrics with Blackbox Differentiation
topic Machine Learning
url https://arxiv.org/abs/1912.03500