PhD Research Seminar: Language-independent word sense disambiguation methods

Мероприятие завершено

Where: Zoom 
When: May 4, 9:30–10:50
Speaker: Maksim Kulaev, second-year PhD student, Faculty of Computer Science

The main idea of word sense disambiguation (WSD) task is to develop an approach for the determination of an appropriate meaning of a word based on its context. A variety of methods have been proposed, from simple dictionary-based techniques to complex neural network systems. However, many of them require a huge amount of labelled data to be trained properly, and they may be limited by a language of an initial dataset.

This talk will be devoted to language-independency in the context of the WSD task. This is quite important because there is a number of languages for which it is crucial due to a lack of data (both structured and unstructured) available. These so-called low-resource languages require specific approaches to obtain sufficient quality. Results of recent experiments will also be presented.