logo SBA

ETD

Archivio digitale delle tesi discusse presso l’Università di Pisa

Tesi etd-04132022-101914


Tipo di tesi
Tesi di laurea magistrale
Autore
MANTOVANI, GIACOMO
URN
etd-04132022-101914
Titolo
Technological troubleshooting supported by sentence embedding with LSTM neural networks
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
ARTIFICIAL INTELLIGENCE AND DATA ENGINEERING
Relatori
relatore Cimino, Mario Giovanni Cosimo Antonio
relatore Vaglini, Gigliola
relatore Bracaloni, Simone
Parole chiave
  • chatbot
  • deep learning
  • long short term memory
  • natural language processing
  • neural networks
  • sentence embedding
  • text classification
Data inizio appello
29/04/2022
Consultabilità
Non consultabile
Data di rilascio
29/04/2092
Riassunto
Organizations looking to increase service productivity may adopt artificial intelligence for time savings and efficiency. Developers build modern chatbots on AI technologies, including deep learning, NLP, and machine learning (ML) algorithms. This project aims to build a chatbot, using LSTM neural networks, which can recognize support requests and help the organization to solve the problem of a customer. We will discuss the tools and techniques needed to reach the final goal such as data extraction and analysis, text preprocessing, sentence embeddings, and so on. We will explain some of the most common architecture for deploying the application and make it available to the user. Finally, we introduce the concept of data drift and how to adjust the model to allow correct predictions of new problems that may occur over time.
File