Quantum Neural Network Simulator
Project Overview
This interactive web application simulates a Quantum Neural Network (QNN) for binary classification tasks. It demonstrates how quantum computing principles can be applied to machine learning problems, providing an educational tool for those interested in quantum machine learning.
Key Features
- Interactive Quantum Circuit Visualization: 3D visualization of quantum circuits with qubits and quantum gates using Babylon.js
- Real-time Data Visualization: View your dataset and classification boundaries in an interactive 3D environment
- Customizable Datasets: Choose from different classification problems (Circles, Spiral, XOR) with adjustable noise levels
- Configurable QNN Architecture: Select the number of qubits and layers in your quantum neural network
- Training Visualization: Watch the loss function decrease in real-time as the network learns
- Performance Metrics: View accuracy and other metrics after training
Technical Implementation
The simulator is built entirely with JavaScript and HTML5, using:
- Babylon.js for all 3D visualizations (quantum circuit, data points, and training progress)
- Custom quantum circuit simulation implementing basic quantum gates (Rx, Ry, Rz, CZ)
- Complex number operations for quantum state manipulation
- Gradient-based optimization for training the quantum neural network
Educational Value
This simulator helps users understand:
- How quantum circuits can be structured for machine learning tasks
- The role of entanglement in quantum neural networks
- How classical data can be encoded into quantum states
- The challenges of training quantum machine learning models
No quantum computing hardware or advanced mathematics knowledge is required to use this simulator, making quantum machine learning concepts accessible to a wider audience.
Future Developments
Planned enhancements include:
- Additional quantum gate types
- More complex dataset options
- Visualization of decision boundaries
- Comparison with classical neural networks
- Export/import of trained models
indexNeuron.zip (7.5 KB)
Quantum-Inspired Counter Visualization
This program demonstrates a quantum-inspired approach to counting from 1 to 10. It simulates quantum measurements and displays the results as a 3D histogram.
Why You Can’t Simply Count 1 to 10 on a Quantum Computer?
Quantum computers don’t work like classical computers when it comes to counting sequentially. Here’s why:
Superposition: Quantum bits (qubits) exist in multiple states simultaneously until measured. When you try to count sequentially, you’re forcing definite states, which defeats the quantum advantage.
Probabilistic Nature: Quantum measurements collapse superpositions into classical states with certain probabilities. This program simulates this by using random sampling - each “measurement” has a probability of yielding numbers 1-10.
No Deterministic Sequences: Quantum algorithms excel at parallel processing and probability distributions, not at deterministic sequences like “1, 2, 3…”. When measured, qubits give probabilistic outcomes.
Measurement Destroys Information: Each measurement collapses the quantum state, making sequential operations challenging.
This visualization shows what happens when we take 500 measurements of a 4-qubit system and count how many times each number from 1-10 appears. The 3D bars represent the frequency distribution - a fundamentally quantum concept rather than classical counting.
Quantum computers are powerful for specific tasks like factoring large numbers or searching unsorted databases, but they require completely different algorithmic approaches than classical step-by-step counting.
indexCount1-10.zip (1.6 KB)
