Qubit state (Bloch vector) Target |1⟩ Trajectory this cycle
drag disk to rotate view

AI Qubit Pulse Optimizer 2D — Bloch-Disk Pulse Trainer

Real quantum computers can't rely on a perfect, textbook control pulse — detuning drifts and decoherence corrupt every gate, and this is exactly the gap AI-driven calibration is starting to fill. This 2D simulator shows a segmented control pulse trying to flip a qubit from |0⟩ to |1⟩ on a rotatable flattened projection of the Bloch sphere while random detuning and dephasing noise fight it every cycle. A stochastic hill-climbing optimizer — the same family of zeroth-order methods used for real pulse-level quantum control — perturbs the pulse shape episode by episode, keeping any change that raises fidelity, so you watch the pulse bars reshape themselves and the fidelity learning curve climb in real time as the AI learns to out-run the noise. Drag the disk to orbit the view and inspect the trajectory from any angle.