Institutional Repository of Vilnius University
Private data navigation with localized ai-based chatbot /
Abstract
dc:descriptionTraditional search engines, while effective for general inquiries, fall short in providing direct and context-specific answers. The advancements of large language model development, offer novel alternatives, capable of generating context-aware responses, using natural language processing. This thesis explores the deployment of localized and secure Retrieval Augmented Generation (RAG) systems, leveraging transformer architectures and vector databases. This work describes the RAG structure and deployment options, maintaining the system’s operations independent of cloud platforms. The experiments assess the system’s response to 100 questions, which vary in their contextual complexity, while also considering cost-effectiveness. The results demonstrate that localized RAG systems can provide relevant answers and, as their accuracy continues to improve, have the potential to extend beyond traditional search methods.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rešetnikovas, Andrius,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD210642390&prefLang=en_US