📉 Simulator: Loss Function (Cross Entropy)
L = -log(ŷ_target)
THE IMAGE'S ACTUAL CLASS (GROUND TRUTH: y_c = 1)
PROBABILITIES ESTIMATED BY SOFTMAX (ŷ_c)
🏠 House (ŷ_1):
0.70
😊 Happy (ŷ_2):
0.20
😢 Sad (ŷ_3):
0.10
Penalty Curve L = -log(ŷ_target)
LOSS CALCULATION:
Loss L = 0.3567
Description: Figure 9.7: Interactive simulator of the Loss Function ( Cross-Entropy ): select the real class of the image (House, Happy or Sad) and adjust the probabilities estimated by Softmax to visualize the calculation of the scalar penalty and the graph of the negative logarithm in real time.