Construct an LLM-based chatbot for a public service: The case of the Greek National Catalogue of Services (MITOS)
Φόρτωση...
Ημερομηνία
2024-06-21
Συγγραφείς
Christodoulou, Theofilos
Τίτλος Εφημερίδας
Περιοδικό ISSN
Τίτλος τόμου
Εκδότης
Δικαιώματα
Default License
Άδειες
Παραπομπή
Παραπομπή
Περίληψη
Περίληψη
The dissertation investigates the feasibility of implementing a Large Language Model (LLM) based chatbot designed to interface with "MITOS," the National Register of Administrative Procedures of the Greek Public Sector. This chatbot is envisioned to offer users an intuitive and efficient platform for querying and comprehending the diverse administrative procedures documented within the register. A key aspect of this innovative approach involves the incorporation of GPT-4, a cutting-edge Large Language Model, augmented by a custom database housing embeddings and contexts of the procedures, stored in an Elasticsearch index.
The synergy between GPT-4's advanced natural language understanding and generation capabilities, and Elasticsearch's robust search functionalities, is intended to yield precise, contextually relevant responses to user inquiries. The thesis explores the integration of these technologies with the goal of enhancing user interaction with the administrative procedures catalogued in "MITOS," ultimately contributing to a more streamlined and user-friendly experience within the Greek Public Sector.
Περιγραφή
Λέξεις-κλειδιά
Large language model (LLM), Greek national catalogue of services, MITOS, Chatbot, Public services