Previous topic

Logistic Regression

Next topic

The Code Tracer

This Page

Datastructure and AlgorithmsΒΆ

The mathematical object is a Univariate Taylor Polynomial over Matrices (UTPM)

\[[x]_D = [x_0, \dots, x_{D-1}] = \sum_{d=0}^{D-1} x_d T^d \;,\]

where each \(x_d\) is some array, e.g. a (5,7) array. This mathematical object is described by numpy.ndarray with shape (D,P, 5,7). P>1 triggers a vectorized function evaluation.

All algorithms are implemented in the following fashion:

def add(x_data, y_data, z_data):
    z_data[...] = x_data[...] + y_data[...]

where the inputs x_data,y_data are numpy.ndarray’s and add changes the elements of the numpy.ndarray z. I.e., the algorithms are implemented in a similar way as LAPACK or Fortran functions in general. One can find the UTPM algorithms in algopy/utpm/algorithms.py where they are class functions of a mixin class.

In practice, working with such algorithms is cumbersome. ALGOPY therefore also offers the class algopy.UTPM which is a thin wrapper around the algorithms and provides overloaded functions and operators. The data is saved in the attribute UTPM.data.

The following code shows how to use the algorithms directly and using the syntactic sugar provided by the UTPM class.

import numpy
from algopy import UTPM, dot

D,P,M,N = 2,1,3,4
x_data = 7*numpy.arange(D*P*M*N,dtype=float).reshape((D,P,M,N))
y_data = numpy.arange(D*P*M*N,dtype=float).reshape((D,P,N,M))

# calling algorithms directly
z_data = numpy.zeros((D,P,M,M))
UTPM._dot(x_data, y_data, z_data)

# use UTPM instance
x = UTPM(x_data)
y = UTPM(y_data)
z = dot(x, y)

print('z.data - z_data', z.data - z_data)