수학곳곳AISubject

AI

Deep learning math: softmax turns scores into probabilities that sum to 1, and backpropagation carries the loss gradient from the output back to the input.

Done
23
Draft
0
Planned
10
Map
See the full map →

Reading order

  1. Softmax

    #008Done

    The softmax function turns scores (logits) into probabilities that sum to 1 by dividing each exp(logit) by the sum. Scores (2.0, 1.0, 0.1) become (0.659, 0.242, 0.099). Covers why exp, temperature, the sigmoid, nan, and the derivative.

  2. Backpropagation

    #019Done한국어

  3. Cross-Entropy

    Planned

  4. Attention

    Planned

  5. Sigmoid

    Planned

  6. Top-k and Top-p Sampling

    Planned

  7. Activation Functions

    Planned

  8. Loss Function

    Planned

  9. Linear Regression

    Planned

  10. Vanishing Gradient

    Planned

  11. Automatic Differentiation

    Planned

  12. Softmax Regression

    Planned

  13. Modes of Machine Learning

    #282Done한국어

  14. Perceptrons

    #283Done한국어

  15. XOR and Multilayer Perceptrons

    #284Done한국어

  16. Expert System Rules

    #285Done한국어

  17. Big Data and Bias

    #286Done한국어

  18. Word Sets for Text

    #287Done한국어

  19. Bag of Words

    #288Done한국어

  20. School Model of TF-IDF

    #289Done한국어

  21. Text Similarity

    #290Done한국어

  22. Lexicon-Based Sentiment Analysis

    #291Done한국어

  23. Pixel Matrices

    #292Done한국어

  24. Image Brightness Transforms

    #293Done한국어

  25. Image Transposition

    #294Done한국어

  26. Image Classification with Hamming Distance

    #295Done한국어

  27. Prediction from Relative Frequencies

    #296Done한국어

  28. Prediction with Trend Lines

    #297Done한국어

  29. Univariate Squared Loss

    #298Done한국어

  30. Slope of a Loss Function

    #299Done한국어

  31. Gradient Descent in One Variable

    #300Done한국어

  32. AI Decision-Making Trade-offs

    #301Done한국어

  33. An AI Mathematics Inquiry

    #302Done한국어