Matlab：在神经网络中使用交叉熵

Question

我正在尝试使用交叉熵实现 RNN。以下是我的代码：

net = layrecnet(1:2,10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
[Xs,Xi,Ai,Ts] = preparets(net, featureMatrix, labels);
net = train(net,Xs,Ts,Xi,Ai);
% view(net)
Y = net(Xs,Xi,Ai);
perf = perform(net,Y,Ts);

performParam来自Matlab官方doc。但是，在我执行它之后，我收到一条警告：

Warning: Performance function replaced with squared
error performance. 
> In trainlm>formatNet (line 155)
  In trainlm (line 65)
  In nntraining.setup (line 14)
  In network/train (line 335)

即使我执行 feedforwardnet，我也得到了同样的警告。以下是我的代码。

[x,t] = simplefit_dataset;
net = feedforwardnet(10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
net = train(net,x,t);
view(net)
y = net(x);
perf = perform(net,y,t);

那么如何在我的代码中使用交叉熵？

Answer 1

问题是 trainlm 仅适用于使用 Jacobian Matrix 的损失函数，如文档所述：

This function uses the Jacobian for calculations, which assumes that performance is a mean or sum of squared errors. Therefore, networks trained with this function must use either the mse or sse performance function.

一种解决方案是使用其他训练算法，例如 trainrp 或 trainscg。以下作品：

[x,t] = simplefit_dataset;
net = feedforwardnet(10);
net.performFcn = 'crossentropy';
net.performParam.regularization = 0.1;
net.performParam.normalization = 'none';
net.trainFcn = 'trainrp';
net = train(net,x,t);
view(net)
y = net(x);
perf = perform(net,y,t);

Matlab：在神经网络中使用交叉熵

Matlab: Use cross-entropy in neural network

matlab

recurrent-neural-network