tf.linalg.LinearOperatorAdjoint

LinearOperator representing the adjoint of another operator.

Inherits From: LinearOperator

This operator represents the adjoint of another operator.

# Create a 2 x 2 linear operator.
operator = LinearOperatorFullMatrix([[1 - i., 3.], [0., 1. + i]])
operator_adjoint = LinearOperatorAdjoint(operator)

operator_adjoint.to_dense()
==> [[1. + i, 0.]
     [3., 1 - i]]

operator_adjoint.shape
==> [2, 2]

operator_adjoint.log_abs_determinant()
==> - log(2)

x = ... Shape [2, 4] Tensor
operator_adjoint.matmul(x)
==> Shape [2, 4] Tensor, equal to operator.matmul(x, adjoint=True)

Performance

The performance of LinearOperatorAdjoint depends on the underlying operators performance.

Matrix property hints

This LinearOperator is initialized with boolean flags of the form is_X, for X = non_singular, self_adjoint, positive_definite, square. These have the following meaning:

  • If is_X == True, callers should expect the operator to have the property X. This is a promise that should be fulfilled, but is not a runtime assert. For example, finite floating point precision may result in these pro