22.02.2024 Views

Daniel Voigt Godoy - Deep Learning with PyTorch Step-by-Step A Beginner’s Guide-leanpub

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Bonus Chapter

Feature Space

This chapter is different from the others. We’re not coding anything.

The purpose of this chapter is to illustrate the effect activation functions have on

the feature space.

"What do you mean by 'feature space'?"

The feature space is the n-dimensional space where our features "live." In Chapter 3

we used two features for our binary classification, so the feature space there was

two-dimensional, a plane. In Chapter 4, we used 25 features, so the feature space

was 25-dimensional, a hyper-plane.

Two-Dimensional Feature Space

Let’s forget about hyper-planes (phew!), and go back to the comfy and familiar

world of two dimensions for now.

This is an entirely new dataset; it is not the image dataset

anymore!

Our new, two-dimensional, data has 2,000 points, evenly split into two classes: red

(negative) and blue (positive), which are neatly organized in two distinct parabolas,

as in the figure below.

Figure B.1 - Two-dimensional feature space

Two-Dimensional Feature Space | 331

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!