之前在https://blog.csdn.net/fengbingchun/article/details/124766283 中介绍过深度学习中的优化算法AdaGrad,这里介绍下深度学习的另一种优化算法RMSProp。
RMSProp全称为Root Mean Square Propagation,是一种未发表的自适应学习率方法,由Geoff Hinton提出,是梯度下降优化算法的扩展。如下图所示,截图来自:https://arxiv.org/pdf/1609.04747.pdf
AdaGrad的一个限制是,它可能会在搜索结束时导致每个参数的步长(学习率)非常小,这可能会大大减慢搜索进度,并且可能意味着无法找到最优值。RMSProp和Adadelta都是在同一时间独立开发的,可认为是AdaGrad的扩展,都是为了解决AdaGrad急剧下降的学习率问题。
RMSProp采用了指数加权移动平均(exponentially weighted moving average)。
RMSProp比AdaGrad只多了一个超参数,其作用类似于动量(momentum),其值通常置为0.9。
RMSProp旨在加速优化过程,例如减少达到最优值所需的迭代次数,或提高优化算法的能力,例如获得更好的最终结果。
以下是与AdaGrad不同的代码片段:
1.在原有枚举类Optimizaiton的基础上新增RMSProp:
enum class Optimization {
BGD, // Batch Gradient Descent
SGD, // Stochastic Gradient Descent
MBGD, // Mini-batch Gradient Descent
SGD_Momentum, // SGD with Momentum
AdaGrad, // Adaptive Gradient
RMSProp // Root Mean Square Propagation
};
2.calculate_gradient_descent函数:RMSProp与AdaGrad只有g[j]的计算不同
void LogisticRegression2::calculate_gradient_descent(int start, int end)
{
switch (optim_) {
case Optimization::RMSProp: {
int len = end - start;
std::vector g(feature_length_, 0.);
std::vector z(len, 0), dz(len, 0);
for (int i = start, x = 0; i < end; ++i, ++x) {
z[x] = calculate_z(data_->samples[random_shuffle_[i]]);
dz[x] = calculate_loss_function_derivative(calculate_activation_function(z[x]), data_->labels[random_shuffle_[i]]);
for (int j = 0; j < feature_length_; ++j) {
float dw = data_->samples[random_shuffle_[i]][j] * dz[x];
g[j] = mu_ * g[j] + (1. - mu_) * (dw * dw);
w_[j] = w_[j] - alpha_ * dw / (std::sqrt(g[j]) + eps_);
}
b_ -= (alpha_ * dz[x]);
}
}
break;
case Optimization::AdaGrad: {
int len = end - start;
std::vector g(feature_length_, 0.);
std::vector z(len, 0), dz(len, 0);
for (int i = start, x = 0; i < end; ++i, ++x) {
z[x] = calculate_z(data_->samples[random_shuffle_[i]]);
dz[x] = calculate_loss_function_derivative(calculate_activation_function(z[x]), data_->labels[random_shuffle_[i]]);
for (int j = 0; j < feature_length_; ++j) {
float dw = data_->samples[random_shuffle_[i]][j] * dz[x];
g[j] += dw * dw;
w_[j] = w_[j] - alpha_ * dw / (std::sqrt(g[j]) + eps_);
}
b_ -= (alpha_ * dz[x]);
}
}
break;
case Optimization::SGD_Momentum: {
int len = end - start;
std::vector change(feature_length_, 0.);
std::vector z(len, 0), dz(len, 0);
for (int i = start, x = 0; i < end; ++i, ++x) {
z[x] = calculate_z(data_->samples[random_shuffle_[i]]);
dz[x] = calculate_loss_function_derivative(calculate_activation_function(z[x]), data_->labels[random_shuffle_[i]]);
for (int j = 0; j < feature_length_; ++j) {
float new_change = mu_ * change[j] - alpha_ * (data_->samples[random_shuffle_[i]][j] * dz[x]);
w_[j] += new_change;
change[j] = new_change;
}
b_ -= (alpha_ * dz[x]);
}
}
break;
case Optimization::SGD:
case Optimization::MBGD: {
int len = end - start;
std::vector z(len, 0), dz(len, 0);
for (int i = start, x = 0; i < end; ++i, ++x) {
z[x] = calculate_z(data_->samples[random_shuffle_[i]]);
dz[x] = calculate_loss_function_derivative(calculate_activation_function(z[x]), data_->labels[random_shuffle_[i]]);
for (int j = 0; j < feature_length_; ++j) {
w_[j] = w_[j] - alpha_ * (data_->samples[random_shuffle_[i]][j] * dz[x]);
}
b_ -= (alpha_ * dz[x]);
}
}
break;
case Optimization::BGD:
default: // BGD
std::vector z(m_, 0), dz(m_, 0);
float db = 0.;
std::vector dw(feature_length_, 0.);
for (int i = 0; i < m_; ++i) {
z[i] = calculate_z(data_->samples[i]);
o_[i] = calculate_activation_function(z[i]);
dz[i] = calculate_loss_function_derivative(o_[i], data_->labels[i]);
for (int j = 0; j < feature_length_; ++j) {
dw[j] += data_->samples[i][j] * dz[i]; // dw(i)+=x(i)(j)*dz(i)
}
db += dz[i]; // db+=dz(i)
}
for (int j = 0; j < feature_length_; ++j) {
dw[j] /= m_;
w_[j] -= alpha_ * dw[j];
}
b_ -= alpha_*(db/m_);
}
}
执行结果如下图所示:测试函数为test_logistic_regression2_gradient_descent,多次执行每种配置,最终结果都相同。图像集使用MNIST,其中训练图像总共10000张,0和1各5000张,均来自于训练集;预测图像总共1800张,0和1各900张,均来自于测试集。在它们学习率为0.01及其它配置参数相同的情况下,AdaGrad耗时为17秒,RMSProp耗时为33秒;它们的识别率均为100%。当学习率调整为0.001时,AdaGrad耗时为26秒,RMSProp耗时为19秒;它们的识别率均为100%。
GitHub: https://github.com/fengbingchun/NN_Test