Sigo publicando Machine Learning Yearning:
|
|
Machine Learning Yearning
|
|
Dear friends,
What happens when the data in your training set does not come from the same distribution as your dev/test set? We are acquiring larger and larger datasets to train our cat detectors...erm...learning algorithms.
Sometimes this means living with a training set whose distribution is different from the test set. When is this a good idea, and how can your algorithms still do well on the data distribution you care about? Read this week’s chapters to learn more!
|
|
As CEO of an AI+Mental health company, my friend Alison Darcy over at Woebotreally liked this one: |
|
| If you have a great AI pun, tweet it to me @AndrewYNg using #AIpun. I'll share my favorite in next week's email! |
|
|
|
|
|
|
