October 2021
Intermediate to advanced
344 pages
8h 51m
English

Bad datasets lead to bad models. We’d like to understand our data before we build a model, and then use that understanding to create a useful dataset, one that leads to models that do what we expect them to do. Knowing basic statistics will enable us to do just that.
A statistic is any number that’s calculated from a sample and used to characterize it in some way. In deep learning, when we talk about samples, we’re usually talking about datasets. Maybe the most basic statistic is the arithmetic mean, commonly known as the average. The mean of a dataset is a single-number summary of the dataset.
We’ll see many different statistics in ...
Read now
Unlock full access