📘 Mathematical Explanation Notebook
View the full step-by-step neural network maths in: neural_network_math.ipynb
Or check out my website: https://s4096770-ai-neural-network-app-v9sprf.streamlit.app/
This project builds a neural network from scratch to model the XOR problem, demonstrating how learning occurs through forward propagation, backpropagation, and gradient descent — all implemented from first principles in Python.
It includes both:
- 📗 A Jupyter Notebook explaining the underlying mathematics step by step
- ⚙️ A Streamlit App to interactively visualise the network’s learning process
🚀 Try it yourself: AI Neural Network App
📘 Read the maths explanation: neural_network_math.ipynb
✅ Implements a 3-layer neural network (2-3-1)
✅ Uses the sigmoid activation and mean squared error loss
✅ Includes full mathematical derivations
✅ Visualises the training loss curve over epochs
✅ Built with Python, NumPy, Matplotlib, and Streamlit
| Tool | Purpose |
|---|---|
| Python 3.9+ | Core implementation |
| NumPy | Linear algebra + optimisation |
| Matplotlib | Visualising learning progress |
| Streamlit | Interactive web app |
| Jupyter Notebook | Mathematical explanation |
| File | Description |
|---|---|
app.py |
Streamlit app for interactive visualisation |
neural_network_math.ipynb |
Jupyter notebook with maths and theory |
neural_network.py |
Core neural network logic |
requirements.txt |
Dependencies for deployment |
Created as a hands-on demonstration of AI fundamentals and mathematical reasoning,
this project shows the full journey from equations ➜ implementation ➜ visualisation.
👩💻 Author: Maha Laeeq
🎓 Bachelor of Computer Science — RMIT University
📍 Melbourne, Australia