Neural Methods for Word Sense Induction

Student: Machnev Alexey

Supervisor: Ekaterina Artemova

Educational Programme: Mathematics and Mathematical Physics (Master)

Year of Graduation: 2021

Word sense induction is a technique, separating different word senses based on context where a word is used. In contrast of word sense disambiguation task, possible word senses are not known, so a set of senses for given word should be retrieved from a list of contexts with given word. The most popular way of solving this task is to use some pre-trained language model with some language knowledge. One good solution is to generate many substitutions for the given word and cluster results. In contrast, we would try to use language model's embeddings directly. In our work, we research methods of building word embeddings, containing information about word sense from BERT model. Then, we compare described methods with approach, based on measuring similarities between contexts using Sequences Pair Classification models.

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses