Online Multi-Source Domain Adaptation through Gaussian Mixtures and Dataset Dictionary Learning

TL;DR

This paper presents a novel online multi-source domain adaptation method using Gaussian Mixtures and dataset dictionary learning, enabling real-time adaptation and memory representation of streaming target data.

Contribution

It introduces a new online Gaussian Mixture Model fitting approach based on Wasserstein geometry and integrates it with dataset dictionary learning for effective online MSDA.

Findings

Successfully adapts in real-time to target data streams

Demonstrates improved adaptation on Tennessee Eastman Process benchmark

Provides a memory-efficient representation of data streams

Abstract

This paper addresses the challenge of online multi-source domain adaptation (MSDA) in transfer learning, a scenario where one needs to adapt multiple, heterogeneous source domains towards a target domain that comes in a stream. We introduce a novel approach for the online fit of a Gaussian Mixture Model (GMM), based on the Wasserstein geometry of Gaussian measures. We build upon this method and recent developments in dataset dictionary learning for proposing a novel strategy in online MSDA. Experiments on the challenging Tennessee Eastman Process benchmark demonstrate that our approach is able to adapt \emph{on the fly} to the stream of target domain data. Furthermore, our online GMM serves as a memory, representing the whole stream of data.