This thesis presents the development of an artificial intelligence-based system for the automatic extraction of travel-related information from natural language texts. The goal is to simplify travel planning through a more intuitive and natural interaction between the user and a virtual assistant. The project is based on the use of NuEstract 2.0, a semantic model capable of identifying and structuring data such as destinations, dates, preferences, or budgets, without altering the original content provided by the user. The main focus is on the design of the extraction system and its integration into a conversational interface, while the analysis of AI models plays a complementary role. The quality of the extraction was assessed through empirical testing aimed at evaluating the system’s accuracy, consistency, and usefulness in realistic scenarios. This project fits within the context of intelligent conversational applications and aims to demonstrate the potential of semantic models in understanding natural language for practical purposes.
Questa tesi presenta lo sviluppo di un sistema basato sull’intelligenza artificiale per l’estrazione automatica di informazioni turistiche da testi in linguaggio naturale. L’obiettivo è facilitare la pianificazione dei viaggi attraverso un’interazione più semplice e intuitiva tra l’utente e un assistente virtuale. Il lavoro si basa sull’impiego di NuEstract 2.0, un modello semantico in grado di individuare e strutturare dati come destinazioni, date, preferenze o budget, senza alterare il contenuto originale fornito dall’utente. L’attenzione principale è rivolta alla progettazione del sistema di estrazione e alla sua integrazione in un’interfaccia conversazionale, mentre l’analisi dei modelli AI ha un ruolo complementare. La qualità dell’estrazione è stata valutata attraverso test empirici, volti a misurare accuratezza, coerenza e utilità del sistema in scenari realistici. Il progetto si inserisce nel contesto delle applicazioni conversazionali intelligenti e mira a dimostrare il potenziale dei modelli semantici nella comprensione del linguaggio naturale per fini pratici.
Dal linguaggio naturale all’itinerario: un sistema AI per l’estrazione di informazioni turistiche
BENIN, ALESSANDRO
2024/2025
Abstract
This thesis presents the development of an artificial intelligence-based system for the automatic extraction of travel-related information from natural language texts. The goal is to simplify travel planning through a more intuitive and natural interaction between the user and a virtual assistant. The project is based on the use of NuEstract 2.0, a semantic model capable of identifying and structuring data such as destinations, dates, preferences, or budgets, without altering the original content provided by the user. The main focus is on the design of the extraction system and its integration into a conversational interface, while the analysis of AI models plays a complementary role. The quality of the extraction was assessed through empirical testing aimed at evaluating the system’s accuracy, consistency, and usefulness in realistic scenarios. This project fits within the context of intelligent conversational applications and aims to demonstrate the potential of semantic models in understanding natural language for practical purposes.| File | Dimensione | Formato | |
|---|---|---|---|
|
Benin_Alessandro.pdf
accesso aperto
Dimensione
3.33 MB
Formato
Adobe PDF
|
3.33 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/89976