Covers: theory of Equivariant representations

- What is equivariance?
- Why equivariance is important?
- What is the effect of using equivariant representations on models?

Watch the first 3.5 minutes of the video to understand what is equivariance and why it is important.

Fail to play? Open the link directly: https://www.youtube.com/watch?v=Y86rzE4UzKs

European Computer Vision Association

0 comment

Collaborators

- Objectives
- The user of this short list will be able to understand the concepts related to convolution motivation, sparse interactions, parameter sharing and equivariant representations
- Potential Use Cases
- You are looking to understand the mathematical foundations for Deep Learning
- Who is this for ?
- BEGINNERDeep Learning practitioners interested in Mathematical foundations

Click on each of the following **annotated items** to see details.

VIDEO 1. Computer vision - parameter sharing

- What's the idea behind parameter sharing?

11 minutes

VIDEO 2. Learning SO(3) Equivariant Representations with Spherical CNNs

- What is equivariance?
- Why equivariance is important?
- What is the effect of using equivariant representations on models?

4 minutes

BOOK_CHAPTER 3. Motivation

- What are the main ideas for Convolutional Neural Networks (CNN)?
- What are sparse interactions?
- What is sparse interactions?
- What is equivariance?

15 minutes

VIDEO 4. Motivation for convolutional layers

- How can we normalize an input image?
- What is the difference of MLP and Convolution?
- What is convolution?
- What does it mean edge detection, Sharpening and blurring??
- How convolution and correlation can be similar?

8 minutes

0 comment