Cross Language Text Classification via Subspace Co-Regularized Multi-View Learning

TL;DR

This paper introduces a novel multi-view learning method that leverages parallel corpora and subspace co-regularization to improve cross-language text classification, reducing the need for extensive labeled data in each language.

Contribution

The paper proposes a new subspace co-regularized multi-view learning approach for cross-language classification using parallel corpora, outperforming existing methods.

Findings

Consistently outperforms inductive, domain adaptation, and multi-view learning methods.

Effective in leveraging parallel corpora for cross-language transfer.

Improves classification accuracy across multiple language pairs.

Abstract

In many multilingual text classification problems, the documents in different languages often share the same set of categories. To reduce the labeling cost of training a classification model for each individual language, it is important to transfer the label knowledge gained from one language to another language by conducting cross language classification. In this paper we develop a novel subspace co-regularized multi-view learning method for cross language text classification. This method is built on parallel corpora produced by machine translation. It jointly minimizes the training error of each classifier in each language while penalizing the distance between the subspace representations of parallel documents. Our empirical study on a large set of cross language text classification tasks shows the proposed method consistently outperforms a number of inductive methods, domain adaptation methods, and multi-view learning methods.