quant_analytics_flow.analytics package

Submodules

quant_analytics_flow.analytics.constants module

quant_analytics_flow.analytics.matrixanalytics module

square_root_symmetric_matrix(A)

Square root of positive semi-definite matrix

\[Q^T \cdot Q = A\]

using a singular value decomposition

Parameters

A (tensor(shape=(N,N))) – Symmetric 2-dimensional tensor

Returns

Returns square root \(Q\)

Return type

Q (tensor(shape=(N,N)))

quant_analytics_flow.analytics.maxsoft module

phi_smooth(x, y)

max_if(x, y)

hyperbolic(x)

Using the hyperbolic function

\[f(x) = \frac{1}{2} \left(x + \sqrt{1 + x^2} \right)\]

Parameters

x (tensor(shape=(...))) – M-dimensional tensor

Returns

Hyperbolic function

Return type

y (tensor(shape=(…)))

soft_max_hypterbolic(x, eps=1e-08)

Using the hyperbolic function to approximate \(\max(x,0)\)

\[g_(x) = f(x/\epsilon)\cdot \epsilon\]

Parameters

• x (tensor(shape=(...))) – M-dimensional tensor

• eps (float64) – scaling parameter

Returns

Hyperbolic function

Return type

y (tensor(shape=(…)))

quant_analytics_flow.analytics.norminv module

norminv(x)

Module contents