Interactive quantum machine learning lab
Quantum features meet classical learning.
Adjust a parameterized circuit, generate synthetic phase data, and watch an illustrative classifier learn what separates two regimes.
Hybrid phase classifier
Data to decision pipeline
4 qubits · 3 layers · hybridIllustrative learning curves
accuracy and normalized lossConfusion matrix
held-out synthetic setAt initialization, predictions are close to chance. Train the illustrative optimizer to expose the tradeoff between feature capacity and noise.
What earns the accuracy?
Ablations isolate which ingredients help under the lab's assumptions. They are comparative teaching aids, not claims about the referenced research.
Ablation studio
Compare four variants using the same synthetic dataset.
Validation accuracy
same split, illustrative scoreOne prediction, unpacked.
Follow four Z expectation values through a small classical head. The arithmetic is simplified so the role of each stage stays inspectable.
Worked example
Step 1 · Quantum readout
The circuit maps a synthetic sample into four bounded expectation values. Each number lies between -1 and +1.
Waiting for step 3
Not evaluatedCheck the mental model.
Five questions distinguish quantum feature extraction, classical learning, device noise, and the limits of this simulation.