Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Issue with dropping terminal measurements for simulation #6759

Open
glanzz opened this issue Oct 9, 2024 · 1 comment
Open

Issue with dropping terminal measurements for simulation #6759

glanzz opened this issue Oct 9, 2024 · 1 comment
Assignees
Labels
kind/bug-report Something doesn't seem to work. triage/needs-more-evidence [feature requests] this seems plausible but maintainers are not convinced about the use cases yet

Comments

@glanzz
Copy link

glanzz commented Oct 9, 2024

Description of the issue
The drop terminal measurements (https://quantumai.google/reference/python/cirq/drop_terminal_measurements) allows user to drop the terminal measures before simulation to get the proper simulation results.
I am trying to optimize a circuit and drop the terminal measurements, but the fidelity and final vector of the optimized circuit is different.
Seems like the measurement operation is absorbing the Z operation even though the Z gate is not separate (part of PhasedXZ gate).
Expected it to replace back the terminal measures like it does for X gate.
Eg:

Optimized:
1: ───PhX(-0.5)^-0.5───!M───

After dropping terminal measures:
1: ───PhX(-0.5)^-0.5───X───

How to reproduce the issue

import cirq as c
import numpy as np



def optimize_circuit(circuit):
    optimized_circuit = c.eject_z(circuit)
    optimized_circuit = c.eject_phased_paulis(optimized_circuit)
    c.testing.assert_circuits_with_terminal_measurements_are_equivalent(
        circuit, optimized_circuit
    )

    return optimized_circuit

qubit = c.LineQubit(1)
circuit = c.Circuit()
circuit.append(c.ZPowGate(exponent=0.71)(qubit))
circuit.append(c.Rz(rads=np.pi*0.125)(qubit))
circuit.append(c.PhasedXPowGate(**{"phase_exponent": 0.409, "exponent": 0.248})(qubit))
circuit.append(c.PhasedXZGate(**{"axis_phase_exponent": 0.555, "x_exponent": 0.803, "z_exponent": 0.0358})(qubit))
circuit.append(c.XPowGate(exponent=0.012)(qubit))
circuit.append(c.S(qubit))
circuit.append(c.Y(qubit))
circuit.append(c.PhasedXZGate(**{"x_exponent": 0.452, "z_exponent": 0.0275, "axis_phase_exponent": 0.489})(qubit))
circuit.append(c.Rx(rads=np.pi*0.333)(qubit))
circuit.append(c.Rz(rads=np.pi*0.25)(qubit))
circuit.append(c.Rx(rads=np.pi)(qubit))
circuit += c.inverse(circuit)
circuit.insert(0, [c.H(qubit)], c.InsertStrategy.NEW)

# Copy circuits for with and without measurements
circuit_without_measurement = circuit.copy()

circuit_with_measurement = circuit.copy()
circuit_with_measurement.append(c.measure_each(qubit))


# Transpile circuit which as measurement appended
transpiled_circuit_with_measurement = c.optimize_for_target_gateset(circuit_with_measurement, gateset=c.SqrtIswapTargetGateset())
print("Transpiled circuit with measurements after gateset SqrtIswap & then optimization:")
print(transpiled_circuit_with_measurement)
transpiled_circuit_with_measurement = optimize_circuit(transpiled_circuit_with_measurement)
print(transpiled_circuit_with_measurement)

# Transpile circuit
transpiled_circuit_without_measurement = c.optimize_for_target_gateset(circuit_without_measurement, gateset=c.SqrtIswapTargetGateset())
transpiled_circuit_without_measurement = optimize_circuit(transpiled_circuit_without_measurement)
print("Transpiled circuit without measurements:")
print(transpiled_circuit_without_measurement)


hqubit = c.LineQubit(0)
h_circuit = c.Circuit()
h_circuit.append(c.H(hqubit))


s = c.Simulator()

transpiled_circuit_with_measurement_but_dropped_measures = c.drop_terminal_measurements(transpiled_circuit_with_measurement)
print("Transpiled circuit with measurement but dropped:")
print(transpiled_circuit_with_measurement_but_dropped_measures)

h_circuit_STATE = s.simulate(h_circuit, qubit_order=[hqubit]).final_state_vector
transpiled_circuit_without_measurement_STATE = s.simulate(transpiled_circuit_without_measurement, qubit_order=[q for q in transpiled_circuit_without_measurement.all_qubits()]).final_state_vector
transpiled_circuit_with_measurement_but_dropped_measures_STATE = s.simulate(transpiled_circuit_with_measurement_but_dropped_measures, qubit_order=[q for q in transpiled_circuit_with_measurement_but_dropped_measures.all_qubits()]).final_state_vector

actual_without_measurement_circuit_fidelity = c.fidelity(h_circuit_STATE, transpiled_circuit_without_measurement_STATE)
dropped_measurement_circuit_fidelity = c.fidelity(h_circuit_STATE, transpiled_circuit_with_measurement_but_dropped_measures_STATE)


print("Fidelity of the circuits should be equal and close to 1 with H circuit (w/o measure, dropped):")
print(actual_without_measurement_circuit_fidelity)
print(dropped_measurement_circuit_fidelity)

print("Simulated States of the circuits (w/o measure, dropped):")
print(transpiled_circuit_without_measurement_STATE)
print(transpiled_circuit_with_measurement_but_dropped_measures_STATE)

Output:

Transpiled circuit with measurements after gateset SqrtIswap & then optimization:
1: ───PhXZ(a=0.5,x=-0.5,z=1.0)───M───
1: ───PhXZ(a=0.5,x=-0.5,z=0)───M───
Transpiled circuit without measurements:
1: ───PhXZ(a=0.5,x=-0.5,z=1.0)───
Transpiled circuit with measurement but dropped:
1: ───PhXZ(a=0.5,x=-0.5,z=0)───I───
Fidelity of the circuits should be equal and close to 1 with H circuit (w/o measure, dropped):
0.999999880790714
0.0

Cirq version: 1.4.0

@glanzz glanzz added the kind/bug-report Something doesn't seem to work. label Oct 9, 2024
@NoureldinYosri NoureldinYosri added the triage/discuss Needs decision / discussion, bring these up during Cirq Cynque label Oct 15, 2024
@NoureldinYosri NoureldinYosri added triage/needs-more-evidence [feature requests] this seems plausible but maintainers are not convinced about the use cases yet and removed triage/discuss Needs decision / discussion, bring these up during Cirq Cynque labels Oct 16, 2024
Copy link

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 30 days

@mhucka mhucka added T status/stale This has been closed due to inactivity for an extended period of time. triage/stale and removed triage/stale labels Nov 18, 2024
@github-actions github-actions bot removed the status/stale This has been closed due to inactivity for an extended period of time. label Nov 19, 2024
@mhucka mhucka removed the T label Nov 19, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
kind/bug-report Something doesn't seem to work. triage/needs-more-evidence [feature requests] this seems plausible but maintainers are not convinced about the use cases yet
Projects
None yet
Development

No branches or pull requests

4 participants