Generalization curve

A plot of both training loss and validation loss as a function of the number of iterations. A generalization curve can help you detect possible overfitting.¹

Footnotes

developers.google.com/machine-learning/glossary#generalization_curve

Loss curve “A plot of loss as a function of the number of training iterations.”
Generalization (machine learning) “To encourage generalization, regularization helps a model train less exactly to the peculiarities of the data i…”
ML crash course - Linear regression “When training a model, you'll often look at a loss curve to determine if the model has converged.”
A simple way to prevent neural networks from overfitting “However, overfitting is a serious problem in such networks.”
Early stopping “In early stopping, you intentionally stop training the model when the loss on a validation dataset starts to in…”
ML crash course - Logistic regression “Logistic regression models use log loss as the loss function instead of squared loss.”
Test set “The training set and validation set are both closely tied to training a model.”
Test loss “A large gap between test loss and training loss or validation loss sometimes suggests that you need to increase…”
Area under the ROC curve 주제어가 겹침
ROC curve “The shape of an ROC curve suggests a binary classification model's ability to separate positive classes from ne…”
Validation set “For example, a single example shouldn't belong to both the training set and the validation set.”
mini-batch “Therefore, each iteration determines the loss on a random 20 of the 1,000 examples and then adjusts the weights…”
손실 함수 “The goal of training is typically to minimize the loss that a loss function returns.”
Log loss 관련 절: Rationale
ML crash course - Classification “If you want to evaluate a model's quality across all possible thresholds, you need ROC curve and AUC.”
Self-training (machine learning) “Notice that each iteration of Step 2 adds more labeled examples for Step 1 to train on.”
Epoch (machine learning) “An epoch represents N/batch size training iterations, where N is the total number of examples.”
L1 loss “L1 loss is less sensitive to outliers than L2 loss.”
Regularization (machine learning) 관련 절: Training loss vs. real-world performance
기계학습 관련 절: Generalization and OOD