Latte: Collaborative Test-Time Adaptation of Vision-Language Models in Federated Learning

TL;DR

Latte is a federated learning framework that enables vision-language models to adapt to diverse data distributions by maintaining local and external memories, improving performance in decentralized test-time adaptation scenarios.

Contribution

Latte introduces a novel memory-based test-time adaptation framework for federated learning, allowing personalized and robust adaptation across clients with limited data.

Findings

Outperforms existing methods on domain adaptation benchmarks

Maintains high performance with minimal communication overhead

Effectively leverages client similarities for personalized adaptation

Abstract

Test-time adaptation with pre-trained vision-language models has gained increasing attention for addressing distribution shifts during testing. Among these approaches, memory-based algorithms stand out due to their training-free nature and ability to leverage historical test data. However, existing test-time adaptation methods are typically designed for a single domain with abundant data. In decentralized settings such as federated learning, applying these methods individually to each client suffers from limited test data, while directly sharing a single global memory via the server prevents proper personalization to each client's unique distribution. To address this, we propose Latte, a novel framework where each client maintains a local memory to store embeddings from its own historical test data and an external memory to store class prototypes from other relevant clients. During communication, each client retrieves prototypes from similar clients under the server's coordination to expand its memory. For local adaptation, Latte utilizes both embedding similarity and uncertainty to enhance model performance. Our theoretical analysis shows that Latte effectively leverages in-distribution clients while remaining robust to out-of-distribution clients. Extensive experiments on domain adaptation and corruption benchmarks validate that Latte achieves superior performance in decentralized settings, while introducing only negligible communication and computation costs. Our code is available at https://github.com/baowenxuan/Latte .