Merge pull request #155 from kmckiern/jordan_docs

docs for jordan gradient

docs/jordan_gradient.rst

@@ -0,0 +1,21 @@
+Jordan Gradient
+==========================
+
+Overview
+--------
+
+This is an implementation of Stephen Jordan's "Fast Quantum Algorithm for
+Numerical Gradient Estimation" [1]_. Jordan's algorithm utilizes an oracle to sample some
+function at some distance, h, from the point of evaluation. 
+These samples are then `kicked back` to output register via modular addition.
+A quantum Fourier transform (QFT) is then used to transform the output register and
+recover the oracle outputs. To first approximation, this can be used to
+estimate the gradient of the function of interest.
+
+In our implementation, the ``phase_estimation`` algorithm is used to perform the phase kickback and QFT
+transformation. Note that this is computationally more
+expensive than the original formulation by Jordan.
+
+See the ``examples`` folder, ``JordanGradient.ipynb`` for more details on using this module.
+
+.. [1] https://doi.org/10.1103/PhysRevLett.95.050501