🎓 Knowledge-Augmented Academic Assistant (AI Module)

This repository contains the AI core of a Knowledge-Augmented Academic Assistant designed to help students explore universities, get guidance, and receive accurate academic information.

The system is built using a Retrieval-Augmented Generation (RAG) approach, combining:

• 📊 Structured university data
• 🤖 Large Language Models (OpenAI)
• 🎙️ Voice input + transcription pipeline

---

🌐 Live Website: Visit clasia.net

---

🚀 Key Features

• 🤖 AI Chatbot for Students

  • Provides guidance on universities, programs, and decisions
  • Answers contextual and follow-up questions

• 📚 RAG-Based Knowledge System

  • Retrieves relevant university data
  • Augments LLM responses with real information
  • Improves factual accuracy and reduces hallucinations

• 🎙️ Voice Input Support

  • Accepts audio queries
  • Converts speech → text → AI response

• 🧠 Context-Aware Conversations

  • Maintains chat history
  • Generates more relevant responses over time

---

🧠 RAG Architecture (Core Idea)

This project follows a Retrieval-Augmented Generation pipeline, which works in two main stages:

1. Retrieval Stage

• User query is processed
• Relevant university data is selected from the internal dataset
• Context is prepared dynamically

2. Generation Stage

• Retrieved context + user query is sent to the LLM
• OpenAI model generates a grounded response

This approach improves reliability because:

• The model does not rely only on pre-trained knowledge
• It uses real, up-to-date, domain-specific data

RAG systems enhance LLM outputs by injecting external knowledge before generation, improving accuracy and reducing hallucination. :contentReference[oaicite:0]{index=0}

---

⚙️ How the System Works

Text Flow

1. User sends a query
2. System checks:
   • Chat history
   • University dataset
3. Relevant context is retrieved
4. OpenAI generates a structured response

Voice Flow

1. User provides audio input (.mp3)
2. Audio is processed and transcribed
3. Transcribed text follows the same RAG pipeline
4. Final response is generated

---

🗂️ Project Structure

├── main.py                       # Core RAG chatbot pipeline (text)
├── modified_main.py              # Enhanced / structured response version
├── audio_main.py                 # Voice input + transcription + RAG pipeline
├── output/                       # Generated outputs and processed data
├── dummy1.mp3 - dummy5.mp3       # Sample audio files for testing

---

---

🧩 Tech Stack

• Python
• OpenAI API (LLM + reasoning)
• Pydantic (structured outputs)
• Audio processing (local handling)
• RAG architecture (custom implementation)

Note: This project primarily uses OpenAI for intelligence. No external speech API dependency is strictly required in this version.

---

🛠️ Setup Instructions

1. Clone the repository

git clone https://github.com/BusraRafa/Knowledge-Augmented-Academic-Assistant.git
cd Knowledge-Augmented-Academic-Assistant

2. Install dependencies

pip install -r requirements.txt

(If requirements.txt is missing, install manually: openai, python-dotenv, pydantic, pydub)

3. Set environment variables

Create a .env file:

OPENAI_API_KEY=your_openai_api_key

▶️ Usage

Run Text-Based Chatbot

python main.py

Run Voice-Based Assistant

python audio_main.py

🎧 Sample Audio

You can test the system using provided files:

• dummy1.mp3
• dummy2.mp3
• dummy3.mp3
• dummy4.mp3
• dummy5.mp3

📌 Notes

• This repository only contains the AI module
• Frontend, UI/UX, and full application integration are handled separately
• Designed to be integrated into:
  • Web apps
  • WhatsApp bots
  • Student platforms

💡 Future Improvements

• Better retrieval optimization (vector DB / embeddings)
• Multilingual support
• Better personalization for students
• Integration with live university APIs
• Text-to-Speech (voice output)