Domain Adaptation for Time-Series Classification to Mitigate Covariate Shift

TL;DR

This paper introduces a supervised domain adaptation method for time-series classification that uses optimal transport and embedding similarity to reduce domain discrepancy and improve performance across different data domains.

Contribution

It proposes a novel two-step supervised domain adaptation approach combining optimal transport and embedding similarity for time-series data.

Findings

Effective in reducing domain shift in time-series classification

Improves accuracy on online handwriting datasets with domain discrepancy

Demonstrates robustness across simulated and real-world datasets

Abstract

The performance of a machine learning model degrades when it is applied to data from a similar but different domain than the data it has initially been trained on. To mitigate this domain shift problem, domain adaptation (DA) techniques search for an optimal transformation that converts the (current) input data from a source domain to a target domain to learn a domain-invariant representation that reduces domain discrepancy. This paper proposes a novel supervised DA based on two steps. First, we search for an optimal class-dependent transformation from the source to the target domain from a few samples. We consider optimal transport methods such as the earth mover's distance, Sinkhorn transport and correlation alignment. Second, we use embedding similarity techniques to select the corresponding transformation at inference. We use correlation metrics and higher-order moment matching techniques. We conduct an extensive evaluation on time-series datasets with domain shift including simulated and various online handwriting datasets to demonstrate the performance.