State-vector simulator (quantum_mind.quantum)

A small, exact NumPy state-vector simulator. It is infrastructure, not a model: the Bloch-sphere viewer uses it to draw trajectories, and the circuits that belong to the quantum-like families (see quantum_mind.circuits) can be simulated and checked with it before they are sent to a back end.

Circuit takes the gates h x y z s sdg rx ry rz p cx cz swap rzz, unitary and diagonal. Angles are plain numbers, trainable weights W(k), input features X(j) or products XX(i, j); a batch of inputs is simulated at once. Circuit.value_and_grad returns adjoint-method gradients (one forward and one backward pass whatever the number of weights), and Circuit.to_qiskit() exports the circuit so it runs unchanged on Aer, IBM Quantum and Amazon Braket through quantum_mind.circuits.run.

from quantum_mind.quantum import Circuit, W

c = Circuit(2)
c.h(0).cx(0, 1).ry(W(0), 1)             # a Bell pair with a trainable rotation
state = c.state([0.4])                  # exact state vector
qc = c.to_qiskit([0.4])                 # the same circuit in Qiskit

Sizes. Up to 22 qubits are accepted. Cost grows as 2ⁿ.

Verification. The simulator is tested gate by gate against Qiskit’s Statevector, and the adjoint gradients against finite differences (tests/test_quantum.py, tests/test_robustness.py).

What is deliberately not here. General-purpose quantum algorithms (VQE, QAOA, Grover, quantum kernels, variational classifiers, the quantum Fourier transform, amplitude estimation) are provided, maintained and benchmarked far better by Qiskit and PennyLane. Version 3.0 removed the copies this library used to carry so that it contains only what is specific to quantum cognition in robotics.