FastFill: Efficient Compatible Model Update

by   Florian Jaeckle, et al.

In many retrieval systems the original high dimensional data (e.g., images) is mapped to a lower dimensional feature through a learned embedding model. The task of retrieving the most similar data from a gallery set to a given query data is performed through a similarity comparison on features. When the embedding model is updated, it might produce features that are not comparable/compatible with features already in the gallery computed with the old model. Subsequently, all features in the gallery need to be re-computed using the new embedding model – a computationally expensive process called backfilling. Recently, compatible representation learning methods have been proposed to avoid backfilling. Despite their relative success, there is an inherent trade-off between the new model performance and its compatibility with the old model. In this work, we introduce FastFill: a compatible model update process using feature alignment and policy based partial backfilling to promptly elevate retrieval performance. We show that previous backfilling strategies suffer from decreased performance and demonstrate the importance of both the training objective and the ordering in online partial backfilling. We propose a new training method for feature alignment between old and new embedding models using uncertainty estimation. Compared to previous works, we obtain significantly improved backfilling results on a variety of datasets: mAP on ImageNet (+4.4%), Places-365 (+2.7%), and VGG-Face2 (+1.3%). Further, we demonstrate that when updating a biased model with FastFill, the minority subgroup accuracy gap promptly vanishes with a small fraction of partial backfilling.


Forward Compatible Training for Representation Learning

In visual retrieval systems, updating the embedding model requires recom...

Towards Universal Backward-Compatible Representation Learning

Conventional model upgrades for visual search systems require offline re...

Neighborhood Consensus Contrastive Learning for Backward-Compatible Representation

In object re-identification (ReID), the development of deep learning tec...

Dual-Tuning: Joint Prototype Transfer and Structure Regularization for Compatible Feature Learning

Visual retrieval system faces frequent model update and deployment. It i...

L^2R: Lifelong Learning for First-stage Retrieval with Backward-Compatible Representations

First-stage retrieval is a critical task that aims to retrieve relevant ...

Dynamic Gradient Reactivation for Backward Compatible Person Re-identification

We study the backward compatible problem for person re-identification (R...

CoReS: Compatible Representations via Stationarity

In this paper, we propose a novel method to learn internal feature repre...

Please sign up or login with your details

Forgot password? Click here to reset