Cross-lingual Approach to Sentiment Analysis

Promotor(s): Els Lefever

Despite the great progress witnessed in recent years for various NLP tasks, lower-resourced languages often have difficulties keeping up because of data scarcity. A possible solution to this problem is the use of cross-lingual information, where knowledge or data from a rich-resourced language, like English, is used to improve the modeling in a low(er)-resourced target language, like Hindi or Dutch. The proposed research aims at constructing a manually labeled data set for the chosen target language, and to develop a cross-lingual Machine Learning approach and evaluate it on the downstream task of Sentiment Analysis (or hate speech detection).

Machine-translation assisted reading

Promotor(s): Lieve Macken

The main goal of this project is to examine whether and how Machine Translation (MT) can be used to support (language) teachers to adequately address linguistic diversity in the classroom and more specifically whether MT can be used as support tool for text comprehension.