neural networks research group
areas
•
demos
•
people
•
projects
•
publications
•
software
•
Nate Kohl
Nate's research focuses on complexification in evolution, and applying learning to real-world tasks like collision warning and robotics.
Email:
nate [at] cs utexas edu
Homepage:
http://www.cs.utexas.edu/users/nate
Publications
Evolving Neural Networks for Strategic Decision-Making Problems
(2009)
Evolving Neural Networks for Fractured Domains
(2008)
Autonomous Learning of Stable Quadruped Locomotion
(2007)
Evolving a Real-World Vehicle Warning System
(2006)
From Pixels to Multi-Robot Decision-Making: A Study in Uncertainty
(2006)
The UT Austin Villa 2006 RoboCup Four-Legged Team
(2006)
Neuroevolution of an Automobile Crash Warning System
(2005)
Evolving Keepaway Soccer Players through Task Decomposition
(2005)
Automatic Feature Selection via Neuroevolution
(2005)
The UT Austin Villa 2005 RoboCup Four-Legged Team
(2005)
Machine Learning for Fast Quadrupedal Locomotion
(2004)
Policy Gradient Reinforcement Learning for Fast Quadrupedal Locomotion
(2004)
The UT Austin Villa 2004 RoboCup Four-Legged Team: Coming of Age
(2004)
The UT Austin Villa 2003 Four-Legged Team
(2004)
UT Austin Villa 2003: A New RoboCup Four-Legged Team
(2003)
Projects
Constructing Intelligent Agents in Simulated Worlds
NEAT: Evolving Vehicle Warning Systems
Cooperative Coevolution of Multi-Agent Systems
NEAT: Evolving Increasingly Complex Neural Network Topologies
Areas of Interest
Neuroevolution
Robotics
Demos
Learning in Fractured Domains