A Learning Chatbot Powered by RAG Technology
Meet your own AI Studymate!
Upload your documents, and a more accurate and secure personalized chatbot is born
- Project Overview
- Architecture
- Demo Video
- My Contributions
- Course Repository
- Team Members
- Project Duration
Chatbots often confuse users by producing false responses. This is also a reason why users are reluctant to use chatbots for learning. So we propose a RAG-based chatbot service named "Perfect StudyMate" that allows students to experience fewer hallucinations. The user uploads the '.pdf' formatted resources used in courses to the chatroom, and the model embeds the uploaded data and stores it in a database. Then, when the user asks chatbot a question related to the data, the model extracts the context related to the question from the embedding database and reflects it in the answer. At the same time, an interactive conversation is implemented by utilizing the history of conversation between the user and the chatbot. Users can enjoy an improved learning experience by talking to chatbots with less hallucination based on the content of the course material they want.
Role: Backend Development
-
Backend Architecture Design
-
Database Design
-
REST API
-
User
- Signup
- Read user
-
Authentication & Authorization
- Login
- Jwt authentication
-
Chatting Session
- Create session
- Read session
- Update session
- Delete session
-
Message
- Send back response from RAG
- Delete message
- Delete messages
-
File Management
- Upload files
- Read files
- Delete files
https://github.com/SecAI-Lab/SWE3028-Fall-2024/tree/main/Team%20K
- Member 1: Juyong Rhee (Team leader)
- Member 2: Yewon Chun
- Member 3: Jihee Hwang
- Member 4: Jorge Alcorta
- Start Date: 2024-09-02
- End Date: 2024-12-13