Tutorial: Neuroevolution of Intelligent Agents

Risto Miikkulainen
Neural Networks Research Group, UT Austin
and
Cognizant AI Lab

Description

Neuroevolution, or optimization of neural networks through population-based search, allows constructing intelligent agents in domains where gradients are not available and significant exploration is needed to find good solutions. This tutorial introduces participants to the basics of neuroevolution, reviews example application areas, and provides hands-on experience through a colab exercise.

Presentation Materials

Slides (4-up pdf), video

Demos

The slides include numerous demos (i.e. animations). They don't run in the 4-up pdf, however they (and more) are included in the Neuroevolution book demo page.

Neuroevolution Exercise (Colab)

This excercise (authored by Yujin Tang) can be run as a notebook in Google Colab. There are three parts:

(1) Neuroevolution for control
(2) Evolutionary Model Merging
(3) Quality Diversity for Model Merging.

Instructions are given in the notebook.

Further Reading

Book: Neuroevolution: Harnessing Creativity in AI Model Design (MIT Press, forthcoming 2025).
A survey article in Science on neuroevolution in neuroscience.
A short summary article on neuroevolution.
A survey article in Nature Machine Intelligence.
Last modified: Sat Jul 12 16:13:27 CDT 2025