CODA: Temporal Domain Generalization via Concept Drift Simulator

TL;DR

CODA introduces a data-centric, model-agnostic framework that simulates future data using feature correlations to improve temporal domain generalization amidst concept drift in real-world datasets.

Contribution

The paper proposes CODA, a novel framework that uses feature correlation matrices to simulate future data, enabling model-agnostic temporal domain generalization.

Findings

CODA effectively simulates future data with feature correlations.

Models trained with CODA data generalize better across time.

CODA outperforms existing methods in temporal domain adaptation.

Abstract

In real-world applications, machine learning models often become obsolete due to shifts in the joint distribution arising from underlying temporal trends, a phenomenon known as the "concept drift". Existing works propose model-specific strategies to achieve temporal generalization in the near-future domain. However, the diverse characteristics of real-world datasets necessitate customized prediction model architectures. To this end, there is an urgent demand for a model-agnostic temporal domain generalization approach that maintains generality across diverse data modalities and architectures. In this work, we aim to address the concept drift problem from a data-centric perspective to bypass considering the interaction between data and model. Developing such a framework presents non-trivial challenges: (i) existing generative models struggle to generate out-of-distribution future data, and (ii) precisely capturing the temporal trends of joint distribution along chronological source domains is computationally infeasible. To tackle the challenges, we propose the COncept Drift simulAtor (CODA) framework incorporating a predicted feature correlation matrix to simulate future data for model training. Specifically, CODA leverages feature correlations to represent data characteristics at specific time points, thereby circumventing the daunting computational costs. Experimental results demonstrate that using CODA-generated data as training input effectively achieves temporal domain generalization across different model architectures.