An advanced AI-driven study assistant that automatically summarizes educational content and generates high-quality multiple-choice questions for active recall learning.
- AI-Powered Summarization: Fine-tuned T5 transformer model for abstractive summarization of educational content
- Smart Question Generation: GPT-style few-shot prompting for generating contextually relevant questions
- High-Quality Distractors: BERT-based distractor generation pipeline for plausible wrong answers
- Question Quality Ranking: Siamese network trained on Quora duplicate question pairs to ensure question quality
- Adaptive Difficulty: Reinforcement learning-based difficulty adjustment using bandit algorithms
- Student Feedback Loop: Dynamic difficulty adjustment based on student performance
- React
- Vite
- Radix UI
- Tailwind CSS
- Python
- FastAPI
- PyTorch
- Hugging Face Transformers
- Sentence Transformers
- Scikit-learn
- Node.js 16+
- Python 3.8+
- CUDA-capable GPU (recommended for optimal performance)
- Clone the repository:
git clone https://github.com/yourusername/ai-study-assistant.git
cd ai-study-assistant- Install frontend dependencies:
npm install- Install backend dependencies:
cd backend
pip install -r requirements.txt- Start the development servers:
Frontend:
npm run devBackend:
cd backend
uvicorn app.main:app --reloadThe application will be available at:
- Frontend: http://localhost:5173
- Backend API: http://localhost:8000
- Enter your study material in the text area
- Select desired difficulty level
- Click "Summarize" to get an AI-generated summary
- Click "Generate Questions" to create multiple-choice questions
- Answer the questions to help the system learn and adjust difficulty
- Outperforms baseline summarizers (TextRank, LexRank) by >25% on ROUGE and BLEU scores
- High-quality question generation with contextually plausible distractors
- Adaptive difficulty adjustment based on student performance
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License - see the LICENSE file for details.
- Hugging Face for the transformer models
- Quora for the question pair dataset
- The open-source community for various tools and libraries