polito.it
Politecnico di Torino (logo)

Deep learning-based Named Entity Recognition models for information extraction in the Automotive domain

Laura Amoroso

Deep learning-based Named Entity Recognition models for information extraction in the Automotive domain.

Rel. Paolo Garza. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering), 2024

Abstract:

As the utilization of Artificial Intelligence has proliferated in recent years, organizations have also demonstrated a growing interest in the potential applications of deep learning technologies to enhance their operational efficiency. This thesis, conducted within the data science division of JATO Dynamics, a leader company in data-driven intelligence for automotive domain, will investigate the utilization of deep learning-based models for the resolution of a Named Entity Recognition task for information extraction. The objective is to facilitate the incorporation of automotive data into the JATO database, enabling the identification of key concepts and transforming unstructured information into a structured format. Our work commenced with an investigation into the models that were employed prior to the advent of transformers (BiLSTM and CRF). This was followed by a comprehensive analysis of the most significant models that are based on transformer technology, BERT and its variants, to prove their benefits. Furthermore, given the growing potential and spread use of Large Language Models, we validated the utilization of the latest GPT to synthesize labels for additional data, which will then be used for the further training of the selected model. We demonstrated that, although this may be an expensive process, it is a highly reliable technique that can effectively automate a manual operation that is currently performed by company experts.

Relators: Paolo Garza
Academic year: 2024/25
Publication type: Electronic
Number of Pages: 77
Additional Information: Tesi secretata. Fulltext non presente
Subjects:
Corso di laurea: Corso di laurea magistrale in Ingegneria Informatica (Computer Engineering)
Classe di laurea: New organization > Master science > LM-32 - COMPUTER SYSTEMS ENGINEERING
Aziende collaboratrici: Jato Dynamics Italia
URI: http://webthesis.biblio.polito.it/id/eprint/33947
Modify record (reserved for operators) Modify record (reserved for operators)