Generalizable Embeddings with Cross-batch Metric Learning
Gurbuz, Yeti Z., Alatan, A. Aydin
–arXiv.org Artificial Intelligence
Global average pooling (GAP) is a popular component in deep metric learning (DML) for aggregating features. Its effectiveness is often attributed to treating each feature vector as a distinct semantic entity and GAP as a combination of them. Albeit substantiated, such an explanation's algorithmic implications to learn generalizable entities to represent unseen classes, a crucial DML goal, remain unclear. To address this, we formulate GAP as a convex combination of learnable prototypes. We then show that the prototype learning can be expressed as a recursive process fitting a linear predictor to a batch of samples. Building on that perspective, we consider two batches of disjoint classes at each iteration and regularize the learning by expressing the samples of a batch with the prototypes that are fitted to the other batch. We validate our approach on 4 popular DML benchmarks.
arXiv.org Artificial Intelligence
Jul-24-2023
- Country:
- Europe > United Kingdom > England > Cambridgeshire > Cambridge (0.04)
- Genre:
- Research Report (0.40)
- Technology: