ALL-IN-1: Short Text Classification with One Model for All Languages

TL;DR

ALL-IN-1 introduces a straightforward multilingual text classification model using SVMs and multilingual embeddings, achieving top performance without parallel data across four languages.

Contribution

The paper presents a simple, effective multilingual classification approach that does not rely on parallel data, outperforming other models in a shared task.

Findings

Ranked 1st out of 12 teams in the shared task

Effective across four diverse languages

Does not require parallel data for training

Abstract

We present ALL-IN-1, a simple model for multilingual text classification that does not require any parallel data. It is based on a traditional Support Vector Machine classifier exploiting multilingual word embeddings and character n-grams. Our model is simple, easily extendable yet very effective, overall ranking 1st (out of 12 teams) in the IJCNLP 2017 shared task on customer feedback analysis in four languages: English, French, Japanese and Spanish.