Daniel@0: function [g, gdata, gprior] = mlpgrad(net, x, t)
Daniel@0: %MLPGRAD Evaluate gradient of error function for 2-layer network.
Daniel@0: %
Daniel@0: %	Description
Daniel@0: %	G = MLPGRAD(NET, X, T) takes a network data structure NET  together
Daniel@0: %	with a matrix X of input vectors and a matrix T of target vectors,
Daniel@0: %	and evaluates the gradient G of the error function with respect to
Daniel@0: %	the network weights. The error funcion corresponds to the choice of
Daniel@0: %	output unit activation function. Each row of X corresponds to one
Daniel@0: %	input vector and each row of T corresponds to one target vector.
Daniel@0: %
Daniel@0: %	[G, GDATA, GPRIOR] = MLPGRAD(NET, X, T) also returns separately  the
Daniel@0: %	data and prior contributions to the gradient. In the case of multiple
Daniel@0: %	groups in the prior, GPRIOR is a matrix with a row for each group and
Daniel@0: %	a column for each weight parameter.
Daniel@0: %
Daniel@0: %	See also
Daniel@0: %	MLP, MLPPAK, MLPUNPAK, MLPFWD, MLPERR, MLPBKP
Daniel@0: %
Daniel@0: 
Daniel@0: %	Copyright (c) Ian T Nabney (1996-2001)
Daniel@0: 
Daniel@0: % Check arguments for consistency
Daniel@0: errstring = consist(net, 'mlp', x, t);
Daniel@0: if ~isempty(errstring);
Daniel@0:   error(errstring);
Daniel@0: end
Daniel@0: [y, z] = mlpfwd(net, x);
Daniel@0: delout = y - t;
Daniel@0: 
Daniel@0: gdata = mlpbkp(net, x, z, delout);
Daniel@0: 
Daniel@0: [g, gdata, gprior] = gbayes(net, gdata);