The AI Program is made up of six faculty and
approximately 30 graduate students within the Department of Computer Science
with strong collaborative ties to other departments at Colorado State and to
Director: Dr. Darrell Whitley, Genetic Algorithms
The following list is a sample of the research currently being conducted in
the AI program. Visit the participating faculty's individual web pages using
links in the above list to
Applications of Genetic Algorithms to seismic problems, development
of new test problems and comparisons of GAs to other local search
and heuristic search methods.
Combinations of genetic algorithms and neural networks. In
particular, this work looks at neurocontrol applications and the
use of grammar based representations to encode neural network
topology and weights.
Scheduling with Genetic Algorithms. The goal will be to revise
existing GA software for a new scheduling application in the
transportation domain. The scheduler will be evaluated and
compared to 2 existing schedulers.
You can download the Genitor package here.
Reinforcement Learning. Function approximation methods are being
investigated as ways to solve multi-step decision problems using
reinforcement learning. Proofs of stability while learning are being
developed by incorporating reinforcement learning into a theoretical
robust control framework. Applications
include the control of robots and energy systems in buildings.
Recognition of EEG Signals. Neural networks are being used to
extract patterns in EEG signals that indicate which mental task a person
is performing, with the goal of providing a new mode of communication for
Computer Graphics. An adaptive assistant is being developed to speed
the tedious process of hand-tracing the two-dimensional contours of regions of
interest in biomedical images, which are then assembled into three-dimensional
Machine Learning Applications in Bioinformatics.
The main focus of Dr. Ben-Hur's lab is prediction of various properties of
proteins: their function, interaction partners, and remote
homologs. The work involves integrating diverse genomic data including
protein sequence, structure, expression, and functional annotations
using tools of machine learning.
Experimental Methods for Evaluating Planning Systems. Development
of experiment designs, analysis techniques, test cases and methods
of modelling AI planning systems in demanding environments.
Improving the reliability of AI Planning systems. Primary focus is
the design and implementation of a prototype debugging tool for AI
Multi-sensor object recognition. Areas being studied include 3D object to
multi-sensor alignment, model representations conducive to multi-sensor
matching, techniques for finding near optimal matches, and 3D real-time
visualization techniques to monitor alignment and recognition procedures.
Near-optimal geometric matching. Local search, genetic algorithms and
hybrids are being studied as heuristic combinatorial optimization
techniques for solving difficult geometric object recognition problems.
The Artificial Intelligence curriculum at CSU consists of a core of the
Additional courses focussed in various artificial intelligence topics,
such as bioinformatics, are
Information on the Graduate Program in Computer Science at
Colorado State University can be obtained from:
Request for additional information can be made
to Dr. Darrell Whitley. Requests
for copies of papers may be sent to any of the participating faculty. E-mail
This page is maintained by
email@example.com. Questions/comments can
be mailed there.