Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper establishes that stationary representations learned through $d$-Simplex fixed classifiers ensure compatibility in feature representations during model updates. By combining cross-entropy loss with contrastive loss, the authors effectively capture higher-order dependencies while maintaining compatibility constraints, which is crucial for sequential fine-tuning. Extensive experiments demonstrate that this approach not only enhances performance during model updates but also allows for uninterrupted retrieval services, achieving state-of-the-art results.
Stationary representations learned through $d$-Simplex classifiers can significantly boost model performance during updates while ensuring seamless retrieval without reprocessing.
Learning compatible representations aims to learn feature representations that can be used interchangeably over time whenever a model undergoes updates. In this paper, we demonstrate that stationary representations learned by $d$-Simplex fixed classifiers imply compatibility as in its formal definition. This result establishes a foundation for future works and can be directly exploited in practical learning scenarios. We address the challenge of learning compatibility using $d$-Simplex fixed classifiers when the model is sequentially fine-tuned. Learning according to a $d$-Simplex fixed classifier with the cross-entropy loss aligns feature distributions at the first-order statistics. Consequently, it may not fully capture higher-order dependencies in the representation between model updates. To address this issue, we demonstrate that training the model using a $d$-Simplex fixed classifier through a convex combination of the cross-entropy loss and a contrastive loss not only captures higher-order dependencies, but is also equivalent to learning with the cross-entropy under the compatibility constraints. We confirm our findings with extensive experiments also considering a new scenario where a pre-trained model is sequentially fine-tuned and occasionally replaced with an improved model. We show that stationary representations enable uninterrupted retrieval services (without reprocessing gallery images) while improving performance during model updates and replacements, achieving state-of-the-art.