6、神经网络学习入门：热冷学习与梯度下降解析

神经网络学习入门：热冷学习与梯度下降解析

1. 热冷学习

热冷学习可能是最简单的学习形式。以下是在Jupyter笔记本中执行的代码，用于尝试正确预测0.8：

weight = 0.5
input = 0.5
goal_prediction = 0.8
step_amount = 0.001
for iteration in range(1101):
    prediction = input * weight
    error = (prediction - goal_prediction) ** 2
    print("Error:" + str(error) + " Prediction:" + str(prediction))

    up_prediction = input * (weight + step_amount)
    up_error = (goal_prediction - up_prediction) ** 2
    down_prediction = input * (weight - step_amount)
    down_error = (goal_prediction - down_prediction) ** 2
    if(down_error < up_error):
        weight = weight - step_amount

    if(down_error > up_error):
        weight = weight + step_amount

这段代码的逻辑是：
- 先进行一次预测，计