current position:Home>Common loss functions

Common loss functions

2022-08-06 09:32:42Ding Jiaxiong

5. Common loss functions

5.1 Loss Function

Used to measure the difference between the model output and the true value

Also called error function, objective function, cost function

A function that measures the quality of the model

5.2 Classification Tasks

5.2.1 Cross-entropy loss function (multi-classification)

  • softmax loss

  • Most used in multi-classification tasks

  • Calculation method

    • insert image description here
  • Purpose: Minimize the negative value of the logarithm of the predicted probability corresponding to the correct class → that is, maximize the probability value

  • insert image description here

  • The closer the predicted value is to the true value, the smaller the loss function

5.2.2 Cross-entropy loss function (sigmoid) for binary classification tasks

  • insert image description here

  • insert image description here

5.3 Return Mission

5.3.1 MAE loss

  • L1 Loss

    • Use absolute error as distance
  • insert image description here

  • insert image description here

5.3.2 MSE loss

  • L2 Loss

    • Use the square of the error as the distance
  • insert image description here

  • insert image description here

5.3.3 smooth L1 loss

  • insert image description here

  • insert image description here

  • insert image description here

copyright notice
author[Ding Jiaxiong],Please bring the original link to reprint, thank you.
https://en.chowdera.com/2022/218/202208060925358745.html

Random recommended