Multidomain Sentence Encoder Training for Semantic Search

Student: Babakov Nikolay

Supervisor: Anastasiya A. Bonch-Osmolovskaya

Educational Programme: Computational Linguistics (Master)

Final Grade: 8

Year of Graduation: 2020

This work is dedicated to the research of possibilities of text vectorization improvement using information about text structure. The main idea is that the most common text structure types (such as question, answer, article title, etc) have unique features that can be used with text encoder for improving generated vectors quality. In this work, a number of experiments have been performed using state-of-the-art methods of machine learning. The main aim of these experiments was to design such encoder architecture and its training pipeline architecture that allow effectively use information about text structure for text vectors generation.

Full text (added June 4, 2020)

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