Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings

tfq.get_sampled_expectation_op

View source on GitHub

Get a TensorFlow op that will calculate sampled expectation values.

tfq.get_sampled_expectation_op(
    backend=None
)

This function produces a non-differentiable TF op that will calculate batches of expectation values given tensor batches of cirq.Circuits, parameter values, and cirq.PauliSum operators to measure. Expectation is estimated by taking num_samples shots per term in the corresponding PauliSum.

# Simulate circuits with C++. 
my_op = tfq.get_sampled_expectation_op() 
# Prepare some inputs. 
qubit = cirq.GridQubit(0, 0) 
my_symbol = sympy.Symbol('alpha') 
my_circuit_tensor = tfq.convert_to_tensor([ 
    cirq.Circuit(cirq.H(qubit) ** my_symbol) 
]) 
my_values = np.array([[0.123]]) 
my_paulis = tfq.convert_to_tensor([[ 
    3.5 * cirq.X(qubit) - 2.2 * cirq.Y(qubit) 
]]) 
my_num_samples = np.array([[100]]) 
# This op can now be run with: 
output = my_op( 
    my_circuit_tensor, ['alpha'], my_values, my_paulis, my_num_samples) 
output 
tf.Tensor([[0.71530885]], shape=(1, 1), dtype=float32) 

In order to make the op differentiable, a tfq.differentiator object is needed. see tfq.differentiators for more details. Below is a simple example of how to make my_op from the above code block differentiable:

diff = tfq.differentiators.ForwardDifference() 
my_differentiable_op = diff.generate_differentiable_op( 
    analytic_op=my_op 
) 

Args:

  • backend: Python object that specifies what backend this op should use when evaluating circuits. It only accepts cirq.Sampler.

Returns:

A callable with the following signature:


  • programs: tf.Tensor of strings with shape [batch_size] containing the string representations of the circuits to be executed.
  • symbol_names: tf.Tensor of strings with shape [n_params], which is used to specify the order in which the values in symbol_values should be placed inside of the circuits in programs.
  • symbol_values: tf.Tensor of real numbers with shape [batch_size, n_params] specifying parameter values to resolve into the circuits specified by programs, following the ordering dictated by symbol_names.
  • pauli_sums: tf.Tensor of strings with shape [batch_size, n_ops] containing the string representation of the operators that will be used on all of the circuits in the expectation calculations.
  • num_samples: tf.Tensor with n_samples[i][j] is equal to the number of samples to draw in each term of pauli_sums[i][j] when estimating the expectation. It can also be tiled up to the shape of pauli_sums by broadcasting if tf.shape(num_samples)[0] or tf.shape(num_samples)[1] is 1 and the other dimension is the same with that of pauli_sums.

  • Returns: tf.Tensor with shape [batch_size, n_ops] that holds the expectation value for each circuit with each op applied to it (after resolving the corresponding parameters in).