Year: 2025
Type: article
Abstract: This project builds an interactive application where users can upload multiple PDF documents and ask questions about them, offering a dynamic way to explore and retrieve information from large texts. The system processes the PDFs by extracting their text, chunking it into smaller sections, and converting these sections into numerical embeddings using advanced language models. These embeddings are stored in a FAISS vector database, enabling efficient similarity search and fast retrieval of relevant information based on user queries. The project uses Stream lit as the frontend framework to create a user-friendly web app, enabling users to interact with the system, upload PDFs, and receive chatbot responses. Accessed via API, powers the conversational AI, generating responses by creating text embeddings for similarity search, which are stored in FAISS for efficient retrieval. Lang Chain orchestrates the interactions between the AI model, memory, and retrieval systems, while utilities like PyPDF2 extract text from PDFs, and dotenv manages environment variables. The chatbot uses Open AI embeddings for text conversion and Conversation Buffer Memory to maintain context throughout user interactions.
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409945080