The dev and test set should:

If you have 10,000,000 examples, how would you split the train/dev/test set?



Size of dev/test set size

University of Michigan - Ann Arbor

Bias and Variance are two fundamental concepts for Machine Learning, and their intuition is just a little different from what you might have learned in your statistics class. Here the video go through two examples that make these concepts super easy to understand.


https://www.youtube.com/watch?app=desktop&v=EuBBz3bI-aA

Bias and Variance

https://www.coursera.org/learn/deep-neural-network?specialization=deep-learning

Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization

Train/dev set error and bias/variance

When the number of training samples tends to infinity, the changing trend of the model trained on the data set:

How does the bias change

- Get more data for better generalization on unseen data.
- Regularization (L2 Regularization, Dropout, Data Augmentation, etc.). 
...

Learn Before

Related

Learn After