Stationary Representations: Optimally Approximating Compatibility and Implications for Improved Model Replacements

TL;DR

This paper establishes that stationary representations learned by the $d$-Simplex fixed classifier optimally approximate compatibility representations, providing a theoretical foundation and practical benefits for model replacement in retrieval systems.

Contribution

It demonstrates the optimality of stationary representations for compatibility approximation and explores their practical advantages in sequential model fine-tuning and replacement scenarios.

Findings

Stationary representations satisfy the formal compatibility constraints.

They enable seamless model replacement without reprocessing gallery images.

Improved retrieval performance during model updates.

Abstract

Learning compatible representations enables the interchangeable use of semantic features as models are updated over time. This is particularly relevant in search and retrieval systems where it is crucial to avoid reprocessing of the gallery images with the updated model. While recent research has shown promising empirical evidence, there is still a lack of comprehensive theoretical understanding about learning compatible representations. In this paper, we demonstrate that the stationary representations learned by the $d$-Simplex fixed classifier optimally approximate compatibility representation according to the two inequality constraints of its formal definition. This not only establishes a solid foundation for future works in this line of research but also presents implications that can be exploited in practical learning scenarios. An exemplary application is the now-standard practice of downloading and fine-tuning new pre-trained models. Specifically, we show the strengths and critical issues of stationary representations in the case in which a model undergoing sequential fine-tuning is asynchronously replaced by downloading a better-performing model pre-trained elsewhere. Such a representation enables seamless delivery of retrieval service (i.e., no reprocessing of gallery images) and offers improved performance without operational disruptions during model replacement. Code available at: https://github.com/miccunifi/iamcl2r.