One hidden layer Neural Network - Derivatives of activation functions


When you implement back propagation for your NN, you need to compute the slop/derivative of the activation function. Let's take a look at this.

  • Sigmoid activation function
figure-1

 We have:

g^{'}(z)=g(z)(1-g(z))=a(1-a)

Following is the derivation:

\\g^{'}(z)=\frac{d}{dz}g(z)\\= (\frac{1}{1+e^{-z}})^{'}\\=\frac{0\times(1+e^{-z})-1\times(1+e^{-z})^{'}}{(1+e^{-z})^{2}}\\=\frac{e^{-z}}{(1+e^{-z})^{2}}=\frac{1+e^{-z}-1}{(1+e^{-z})^{2}}=\frac{1}{(1+e^{-z})}-\frac{1}{(1+e^{-z})^{2}}\\=g(z)(1-g(z))\\=a(1-a)

  • Tanh activation function
figure-2

 

We have:

g^{'}(z)=1-(tanh(z))^{2}=1-a^{2}

Here's the derivation:

g^{'}(z)=\frac{d}{dz}g(z)\\=(\frac{e^{z}-e^{-z}}{e^{z}+e^{-z}})^{'}\\ =\frac{(e^{z}-e^{-z})^{'}\times(e^{z}+e^{-z})-(e^{z}-e^{-z})(e^{z}+e^{-z})^{'}}{(e^{z}+e^{-z})^{2}}\\ =\frac{(e^{z}+e^{-z})^{2}-(e^{z}-e^{-z})^{2}}{(e^{z}+e^{-z})^{2}}\\ =1-(tanh(z))^{2}\\ =1-a^{2}

  • ReLU activation function
figure-3

We have:

g^{'}(z)=\left\{\begin{matrix} 0 \: \: if z< 0\\ 1 \: \: if\: z\geqslant 0 \end{matrix}\right.

  • Leaky ReLU activation function
figure-4

 

 g^{'}(z)=\left\{\begin{matrix} 0.01 \: \: if z< 0\\ 1\: \: \: \: \: \: \: if\: z\geqslant 0 \end{matrix}\right.

<end>

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值