An Unsupervised Tensor-Based Domain Alignment

TL;DR

This paper introduces an unsupervised tensor-based domain alignment method that optimizes alignment matrices and subspaces on an oblique manifold, improving speed and accuracy in domain adaptation tasks.

Contribution

It presents a novel tensor-based domain alignment algorithm using oblique manifold optimization, enhancing flexibility and robustness over traditional methods.

Findings

Faster domain adaptation conversion speed

Significantly improved classification accuracy

Outperforms current state-of-the-art techniques

Abstract

We propose a tensor-based domain alignment (DA) algorithm designed to align source and target tensors within an invariant subspace through the use of alignment matrices. These matrices along with the subspace undergo iterative optimization of which constraint is on oblique manifold, which offers greater flexibility and adaptability compared to the traditional Stiefel manifold. Moreover, regularization terms defined to preserve the variance of both source and target tensors, ensures robust performance. Our framework is versatile, effectively generalizing existing tensor-based DA methods as special cases. Through extensive experiments, we demonstrate that our approach not only enhances DA conversion speed but also significantly boosts classification accuracy. This positions our method as superior to current state-of-the-art techniques, making it a preferable choice for complex domain adaptation tasks.