Introduction to AI Labs
Explore how AI works by changing settings and comparing results.
Learning a line
Adjust a parameter by hand, then train a line using gradient descent. Compare how different learning rates affect the error.
Start the lab →Runs in your browser · No coding required · English / 한국어
How a prior changes the MAP estimate
Change the Beta prior and the observed tosses. Watch the likelihood, posterior, MLE and MAP update.
Explore the Beta prior →Interactive graph · No coding required · English / 한국어
Image denoising: MLE and MAP
Change the noise level and prior strength. Compare quadratic and total variation (TV) priors, and examine the trade-off between noise removal and preserving detail.
Start the lab →Runs in your browser · No coding required · Activity in English
Regularization
Regularization penalizes large weights. Change its strength in TensorFlow Playground and observe how the decision boundary changes.
Open TensorFlow Playground →Interactive demo · No coding required · English
Choosing a hyperparameter: polynomial degree
Adjust the degree of a regression curve. Compare training and validation errors, choose a model, and check its test error.
Explore model complexity →Interactive graphs · No coding required · English / 한국어