Exploring Variational Auto-encoders VAEs

May 21, 5:00 – 6:00 PM

67
RSVPs

Learn About a New and Exciting Topic in Machine Learning!

Artificial IntelligenceCodecademy Volunteer LedData ScienceMachine LearningMath

About this event

Part 1: The Foundation: Understanding Autoencoders 

Curious about how machines learn to understand, recreate, and even generate entirely new images?

Variational Autoencoders (VAEs) are a special kind of neural network setup. Unlike regular autoencoders that simply compress data, VAEs can also create new data points. This versatility makes them useful in various fields, such as generating images, understanding language, and detecting anomalies.

In our workshop series, we're starting with Autoencoders, the building blocks of VAEs. Why? Understanding autoencoders first lays a strong foundation for delving into the complexities of VAEs later on. So let's dive into autoencoders and pave the way for exploring VAEs in part 2!

This is hosted by Negar who is a researcher in deep learning and quantum computing, focusing on natural language processing. Since 2020, she's also been creating content and organizing events.

Prerequisites:

1 - A good grasp of Python’s functions, classes, and loops

2 - A basic understanding of PyTorch

In this workshop, you’ll explore how Autoencoders work through fun examples like transforming images into compact codes and then reconstructing them flawlessly. You'll learn how these smart algorithms capture the essence of data, making it easier for computers to understand and manipulate visual information.

By the end of the session, you'll walk away with a solid understanding of Autoencoders, their practical applications, and how they pave the way for Variational Autoencoders. Don't miss out on this chance to unravel the mysteries of machine learning – reserve your spot now!

The Agenda:

  • Overview of Autoencoders (~ 20 minutes)
  • Autoencoders and their purpose in machine learning.
  • The architecture of Autoencoders, including the encoder and decoder components.
  • The key applications of Autoencoders
  • Hands-on Implementation (~ 40 minutes)
  • Setting up the development environment (e.g., Jupyter Notebook or Google Colab).
  • Implementing a simple Autoencoder using PyTorch.
  • Experimenting with different hyperparameters
Questions you might have:
    • Q: Will the event contain a lot of mathematics?
      • A: No, just a tiny bit! but don't worry if you're not a math whiz. We'll explain everything in simple terms, focusing more on intuition than complex formulas.
    • Q: Do I need to have a program installed on my computer before joining the session?
      • A: If you're up for it, you can install Jupyter Notebook on your computer. However, you'll also have the option to run the workshop's notebook on Colab, so no worries if you prefer that route.

Part 2: "The Frontier: Building Variational Autoencoders (VAEs)" June, TBD**

Organizers

  • Mariana B

    Codecademy

    Community Manager

  • Jonathan T

    Codecademy

    Head of Community

Contact Us

67