Institutional Repository of Vilnius University
Teisinių tekstų sutraukimas ir konteksto ištraukimas naudojant giliuosius neuroninius tinklus /
Abstract
dc:descriptionThis paper analyses solutions for natural language processing problems in the context of legal documents in Lithuanian language using transformer-type deep neural networks as well as methods visualization solutions. General RoBERTa architectural type natural language processing model was created and trained with legal texts. Next, created model was fine-tuned for classification of legal documents and semantic search. This paper reviews transformer neural networks, transformer-type BERT and RoBERTa model architectures. Then integrated gradient method for information visualization was reviewed. Research has shown that natural language processing model for legal documents in Lithuanian language can be created. This paper shows how the contextual information extracted from created language model can be applied to the classification of legal documents by document class and semantic search. In addition, this paper demonstrates how the integrated gradients method can be used to visualize information of created classification model.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Tvarijonas, Audrius,
- Contributors dc:contributor
-
- Petkevičius, Linas
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD107115547&prefLang=en_US