Development of rephrase system using Nature Language Processing models

Ivan Tsvietkov

Master's thesis

Abstract:

The thesis is focused on development of rephrase system using Nature Language Processing models in Python. The theoretical part starts with origins Artificial Intelligence field. After this it shortly defines Machine learning and Deep learning, and describes what is it neural network, how it works, why neurons need activation function and types of it. The theoretical part goes on to describe main NLP techniques, types of neural networks which are used to process human language like RNN and Transformers and ends with brief description of tools for NLP engineer. The practical part shows to ways of creating rephase system. The first way shows how to change sentence using synonyms with libraries like NLTK and spaCy. The second way shows more sophisticated way, it takes already pre-trained model of transformer and additionally trains it on specific corpus created for rephase tasks.

Keywords

Deep learning, NLTK, Natural language processing, Machine learning, TensorFlow, TextBlob, Pytorch

Jazyk práce: English

Datum vytvoření / odevzdání či podání práce: 14. 4. 2022

Obhajoba závěrečné práce

• Vedúci: Ing. Petr Hanzlík, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme / odbor:
Systems Engineering and Informatics / Informatics