This document asks and answers a few leading questions about Google's Machine Learning Glossary.
How do I propose a new term?
Click the Send Feedback button at the top or bottom of the Machine Learning Glossary. Then, tell us what term(s) are missing.
Can I report a confusing definition?
Absolutely! Click the Send Feedback button and tell us.
Will you email me a response to my feedback?
Unfortunately not, but we really do review all feedback.
Who writes the definitions?
A Google team of technical writers, researchers, and software engineers writes and reviews each definition.
How often do you add new definitions?
We release batches of new terms three to four times per year. Additionally, we frequently make minor changes to existing definitions.
Do you provide subglossaries?
Yes, you can filter the glossary by choosing a topic from the Glossary drop-down in the top navigation bar. For example, the Clustering subglossary focuses on the terms relevant to, well, clustering.
Who is the target audience?
We aim the "Fundamentals" terms (those marked with a 🐣 icon) at machine learning newcomers. At the other extreme, we aim certain advanced terms at experienced machine learning practitioners.
Why does my organization define terms differently than your glossary?
That's a pretty common problem in relatively new fields like machine learning. We'll just offer our apologies and promise that machine learning terminology will probably never be completely standardized.
Why do some definitions read like encyclopedia entries?
A single sentence definition suffices for some terms, but other terms require extensive explanations, images, comparisons, examples, math formulas, and so on.