The thesis describes the development of a web application capable of offering chatbot functionalities by leveraging an LLM model implemented on cloud infrastructures that can interpret and respond to natural language requests, while ensuring high performance and easy accessibility for users. The main objective is to create an interactive and scalable solution, also providing the possibility to utilize locally developed custom LLMs. The work is structured into several phases: the analysis of functional and non-functional requirements, the design of the system architecture, and the implementation of both the backend and the user interface. In particular, the integration of the LLM model is explored in depth, with a focus on managing communication between the front-end and the cloud service, as well as ensuring the security and privacy of user data. Through a series of tests and considerations, the thesis demonstrates how adopting cloud technologies can facilitate the deployment and scalability of the application, making smooth and efficient conversational interaction possible. It also shows the possibility of using a customized LLM for more targeted and appropriate responses. The experimental results highlight the chatbot’s ability to respond in a coherent and contextually relevant manner, opening the door to potential applications in various sectors, such as customer support and virtual assistance. Finally, the work discusses the potential for further developments and optimizations, laying the foundation for future research in the field of conversational artificial intelligence and integrated cloud-based web solutions.
La tesi descrive lo sviluppo di un'applicazione web in grado di offrire funzionalità di chatbot sfruttando un modello LLM implementato su infrastrutture cloud in grado di interpretare e rispondere a richieste in linguaggio naturale, garantendo al contempo prestazioni elevate e una facile accessibilità da parte degli utenti. L'obiettivo principale è quello di creare una soluzione interattiva, scalabile e con anche la possibilità di utilizzare di LLM locali sviluppati su misura. Il lavoro è strutturato in diverse fasi: l'analisi dei requisiti funzionali e non funzionali, la progettazione dell'architettura di sistema e l'implementazione del backend e dell'interfaccia utente. In particolare, viene approfondita l'integrazione del modello LLM, con particolare attenzione alla gestione della comunicazione tra il front-end e il servizio cloud, nonché alla sicurezza e alla privacy dei dati dell'utente. Attraverso una serie di test e considerazioni, la tesi dimostra come l'adozione di tecnologie cloud possa facilitare il deployment e la scalabilità dell'applicazione, rendendo possibile un'interazione conversazionale fluida ed efficiente, dimostrando anche la possibilità di adottare un LLM personalizzato per risposte più mirate ed adeguate. I risultati sperimentali evidenziano la capacità del chatbot di rispondere in modo coerente e contestualmente rilevante, aprendo la strada a possibili applicazioni in diversi settori, quali il supporto al cliente e l'assistenza virtuale. Infine, il lavoro discute le potenzialità di ulteriori sviluppi e ottimizzazioni, ponendo le basi per future ricerche nel campo dell'intelligenza artificiale conversazionale e delle soluzioni web-based integrate su cloud.
Web application con funzionalità di chatbot tramite LLM su cloud
MARANGON, ZACCARIA
2024/2025
Abstract
The thesis describes the development of a web application capable of offering chatbot functionalities by leveraging an LLM model implemented on cloud infrastructures that can interpret and respond to natural language requests, while ensuring high performance and easy accessibility for users. The main objective is to create an interactive and scalable solution, also providing the possibility to utilize locally developed custom LLMs. The work is structured into several phases: the analysis of functional and non-functional requirements, the design of the system architecture, and the implementation of both the backend and the user interface. In particular, the integration of the LLM model is explored in depth, with a focus on managing communication between the front-end and the cloud service, as well as ensuring the security and privacy of user data. Through a series of tests and considerations, the thesis demonstrates how adopting cloud technologies can facilitate the deployment and scalability of the application, making smooth and efficient conversational interaction possible. It also shows the possibility of using a customized LLM for more targeted and appropriate responses. The experimental results highlight the chatbot’s ability to respond in a coherent and contextually relevant manner, opening the door to potential applications in various sectors, such as customer support and virtual assistance. Finally, the work discusses the potential for further developments and optimizations, laying the foundation for future research in the field of conversational artificial intelligence and integrated cloud-based web solutions.| File | Dimensione | Formato | |
|---|---|---|---|
|
Web application with chatbot functionality using a cloud based LLM.pdf
Accesso riservato
Dimensione
1.13 MB
Formato
Adobe PDF
|
1.13 MB | Adobe PDF |
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/20.500.12608/89995