batchDerivative: One iteration of the TAGI with derivative calculations

In mgoulet847/tagi: Tractable Approximate Gaussian Inference in Neural Networks

Description

This function goes through one learning iteration of the neural network model using TAGI with derivative calculations.

Usage

batchDerivative(NN, theta, normStat, states, x, Sx, y, dlayer)

Arguments

NN	Lists the structure of the neural network
theta	List of parameters
normStat	Normalized statistics
states	List of states
x	Input data
Sx	Variance of input data
y	Response data
dlayer	Layer from which derivatives will be in respect to

Value

• List of parameters

• List of normalized statistics

• Mean of predicted responses

• Variance of predicted responses

• Mean of first derivative of predicted responses

• Variance of first derivative of predicted responses

• Covariance between derivatives and inputs

• Mean of second derivative of predicted responses

mgoulet847/tagi documentation built on Dec. 21, 2021, 5:10 p.m.