AI-Driven Real-Time Query Management and FAQ System

Overview

This repository implements an AI-driven query management system for large-scale online learning environments. It leverages state-of-the-art Natural Language Processing (NLP) techniques to handle student queries in real time, automate FAQ generation, and ensure scalability and responsiveness during live sessions.

Key Features

• Semantic Embedding: Uses paraphrase-mpnet-base-v2 for high-dimensional vector representation of course material.
• Topic Clustering: Organizes embedded content with FAISS and k-means clustering for efficient query resolution.
• Automated FAQ Generation: Pre-validates FAQs in various categories (in-chapter, across-chapter, application-based, out-of-context).
• Retrieval-Augmented Generation (RAG): Provides accurate, low-latency responses for routine queries.
• Query Escalation: Escalates high-dissimilarity queries to instructors for personalized feedback.
• Performance Metrics:
  • Answer Relevancy: 88.65%
  • Contextual Precision: 88.25%
  • Hallucination Rate: 74.6%

System Workflow

1. Content Preparation:
   • Upload course material (e.g., textbooks, notes).
   • Embed content and create topic clusters using FAISS indexing.
2. Real-Time Query Handling:
   • Embed student queries and calculate dissimilarity with FAQs.
   • Address low-dissimilarity queries automatically with RAG.
   • Flag high-dissimilarity queries for instructor review.
3. Dynamic Feedback:
   • Provide instructors with a dashboard to monitor unresolved queries.
   • Enable updates to the FAQ database for adaptive learning.

Installation

1. Clone the repository:

   git clone https://github.com/<your-repo-url>.git

2. Install dependencies:

   pip install -r requirements.txt

3. Set up FAISS indexing and upload course material using the provided scripts.

Usage

• Start the query management system:

  python run_query_manager.py

• Access the instructor and student dashboards through the web interface.

🚧 Work in Progress

Current Progress: [██████████████████----------] 60%

Future Enhancements

• Multilingual query handling using models like mBERT and XLM-R.
• Advanced reasoning for interdisciplinary and multi-step queries.
• Real-time content updates for evolving curricula.

---