Documentation
¶
Overview ¶
Package gradient provides quantum gradient computation methods.
CostFunc creates an optim.ObjectiveFunc from a parameterized circuit and Hamiltonian using statevector simulation.
ParameterShift and FiniteDifference compute gradients of an objective function, suitable for use as optim.GradientFunc.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
func CostFunc ¶
func CostFunc(circuit *ir.Circuit, hamiltonian pauli.PauliSum, paramNames []string) optim.ObjectiveFunc
CostFunc creates an ObjectiveFunc that evaluates the expectation value ⟨ψ(θ)|H|ψ(θ)⟩ using statevector simulation. Each call binds the parameter vector to paramNames and runs a fresh simulation.
The returned function is goroutine-safe because each call creates its own simulator instance.
func FiniteDifference ¶
func FiniteDifference(f optim.ObjectiveFunc, h float64) optim.GradientFunc
FiniteDifference returns approximate gradients via central differences with step size h.
func ParameterShift ¶
func ParameterShift(f optim.ObjectiveFunc) optim.GradientFunc
ParameterShift returns exact gradients via the parameter-shift rule. For each parameter, it evaluates f(x + π/2*e_i) and f(x - π/2*e_i), computing the gradient as [f(x+s) - f(x-s)] / 2. This requires 2*N function evaluations per call (N = number of parameters).
Types ¶
This section is empty.
Source Files