English

Piecewise-Linear Manifolds for Deep Metric Learning

Computer Vision and Pattern Recognition 2024-03-25 v1 Artificial Intelligence Machine Learning Image and Video Processing

Abstract

Unsupervised deep metric learning (UDML) focuses on learning a semantic representation space using only unlabeled data. This challenging problem requires accurately estimating the similarity between data points, which is used to supervise a deep network. For this purpose, we propose to model the high-dimensional data manifold using a piecewise-linear approximation, with each low-dimensional linear piece approximating the data manifold in a small neighborhood of a point. These neighborhoods are used to estimate similarity between data points. We empirically show that this similarity estimate correlates better with the ground truth than the similarity estimates of current state-of-the-art techniques. We also show that proxies, commonly used in supervised metric learning, can be used to model the piecewise-linear manifold in an unsupervised setting, helping improve performance. Our method outperforms existing unsupervised metric learning approaches on standard zero-shot image retrieval benchmarks.

Keywords

Cite

@article{arxiv.2403.14977,
  title  = {Piecewise-Linear Manifolds for Deep Metric Learning},
  author = {Shubhang Bhatnagar and Narendra Ahuja},
  journal= {arXiv preprint arXiv:2403.14977},
  year   = {2024}
}

Comments

Accepted at CPAL 2024 (Oral)