🧠 AI-Powered Adaptive Teaching Assistant

An intelligent teaching assistant system that understands student queries and generates personalized learning recommendations using semantic embeddings and adaptive progression logic.

---

📌 Project Overview

This project implements a hybrid AI system that:

• Understands student queries using semantic embeddings (Sentence-BERT)
• Classifies query intent and topic using machine learning
• Predicts difficulty level dynamically
• Models student learning state using historical performance
• Generates personalized next-step recommendations

The system combines NLP-based query understanding with performance-aware adaptive learning.

---

🏗️ System Architecture

The system consists of two major components:

1️⃣ Query Understanding Module

• SentenceTransformer (all-MiniLM-L6-v2) for semantic embeddings
• Logistic Regression for:
  • Intent classification
  • Topic classification
• Rule-based difficulty adjustment

2️⃣ Adaptive Learning Path Module

• Student performance analysis from learning logs
• State modeling (quiz score, attempts, confidence)
• Rule-based recommendation engine for:
  • Revision
  • Advancement
  • Continuation

---

🔍 Key Highlights

• Semantic query understanding using Sentence-BERT embeddings
• Dual classifier system (Intent + Topic)
• Dynamic difficulty adjustment logic
• Student performance-aware progression system
• Modular and scalable architecture

---

📂 Project Structure

ai-teaching-assistant/
│
├── app.py
│
├── data/
│   ├── student_queries.csv
│   └── student_learning_logs.csv
│
├── models/
│   ├── __init__.py
│   ├── difficulty_predictor.py
│   ├── intent_classifier.py
│   ├── topic_classifier.py
│   └── embeddings.py
│
├── learning_path/
│   ├── __init__.py
│   ├── student_state.py
│   └── recommendation_policy.py

---

⚙️ How It Works

1. User submits a learning query.
2. Query is converted into embeddings using Sentence-BERT.
3. Intent and topic are predicted using trained classifiers.
4. Difficulty level is dynamically adjusted.
5. Student performance history is analyzed.
6. A personalized learning recommendation is generated.

---

🚀 How to Run

pip install sentence-transformers scikit-learn pandas
python ai-teaching-assistant/app.py

Query: i dont understand how backpropagation works
Intent: Explanation
Topic: Backpropagation
Difficulty: Intermediate
Next Topic: Backpropagation
Action: Revision
Difficulty Adjustment: Decrease


---