NeuroEvolution of Augmenting Topologies and Compositional Pattern Producing Networks
Offered By: Aleksa Gordić - The AI Epiphany via YouTube
Course Description
Overview
Explore the concepts of NeuroEvolution of Augmenting Topologies (NEAT) and Compositional Pattern Producing Networks (CPPN) in this comprehensive video lecture. Delve into the basic ideas behind NEAT, including genome structure, mutation processes, and speciation techniques. Understand how NEAT evolves both network weights and architectures, and learn about its application to the XOR task. Discover the principles of CPPNs, their role in modeling developmental biology, and their ability to generate complex patterns through function composition. Examine the expressiveness of CPPNs in creating symmetries and repetitions, and gain insights into the HyperNEAT concept. Enhance your knowledge of evolutionary algorithms and their applications in neural network design and pattern generation.
Syllabus
Intro to NEAT and CPPNs
Basic ideas behind NEAT
NEAT genome explained
Competing conventions problem
NEAT mutations explained
NEAT genome mating explained
Maintaining innovations via speciation
Explicit fitness sharing
NEAT on XOR task
CPPNs and neural automata
Spatial signal as a chemical gradient abstraction
Composing functions
CPPN main idea recap
Breeding "images" using CPPNs
CPPNs are highly expressive symmetries, repetition...
HyperNEAT idea explained
Outro
Taught by
Aleksa Gordić - The AI Epiphany
Related Courses
Neural Networks for Machine LearningUniversity of Toronto via Coursera Good Brain, Bad Brain: Basics
University of Birmingham via FutureLearn Statistical Learning with R
Stanford University via edX Machine Learning 1—Supervised Learning
Brown University via Udacity Fundamentals of Neuroscience, Part 2: Neurons and Networks
Harvard University via edX