Computer gaming, which uses sophisticated mathematical models, can be used to come up with decision-making strategies that will benefit government agencies, businesses and other institutions, according to a UTEP professor who received a prestigious grant to help further his research.
Christopher Kiekintveld, 32, a computer science professor at the University of Texas at El Paso, said the $488,288 grant from the National Science Foundation will help him and his research team "study ways to improve the robustness of decision-making in the face of challenging problems."
"My research in the area of artificial intelligence is driven by fundamental questions about how we can use computational analysis techniques to make good decisions in highly complex environments," Kiekintveld said.
"Robustness," in this case, refers to operating systems or other programs that can perform effectively under ordinary and unusual conditions, according to the Linus Information Project's website www.linfo.org.
The Computer Science Department at Stanford University defines artificial intelligence as the science and engineering of making intelligent machines, especially intelligent computer programs, of which some applications include game playing and speech recognition.
Experts said that coming up with the right mathematical model is often the key to developing such breakthroughs.
While Kiekintveld is not involved in developing models for crime investigations,
he agrees that some of his cutting-edge work resembles what viewers have come to see on the now-syndicated CBS television drama "Numb3rs." It's pretty much about getting the right numbers in the right place to arrive at a solution.
"One of the big application areas we have been looking at is using game models to help understand how to allocate and schedule security resources in an efficient way," Kiekintveld said. "One of our primary techniques is the use of computational game theory, which is a mathematical way of representing decision-making problems.
"Games like chess are an example of this," he said. "You are making a choice between different strategies, and the key is that your choice depends on what your opponent, or the other player, is likely to do. So we have to predict what decisions they are likely to make and choose a strategy based on that prediction. We can model all that mathematically, and use computers to analyze the decisions."
Some examples of multi-agent systems, Kiekintveld said, include teams of rescue workers, soldiers, or robots coordinating to achieve a mission; trading agents participating in auctions of financial markets; and security forces trying to prevent attacks on critical
infrastructure or computer networks.
"The unifying theme across these areas is predicting how other intelligent agents will act, based on knowledge of their capabilities, goals, historical patterns of behavior, and other factors," he said.
The idea of smart machines that can predict behaviors that are harmful and can be prevented is portrayed, although on a much larger scale, in "Person of Interest," another CBS TV series.
The grant Kiekintveld received is part of the National Science Foundation's Faculty Early Career Development (CAREER) program, which offers prestigious awards to junior faculty members for work that blends education and research.
Both graduate and undergraduate students will take part in the project, by designing algorithms and "agents" to play in tournaments. The agents will be used to help evaluate the robustness of different strategic reasoning methods.
"The award that Dr. Kiekintveld received is called the CAREER award because it is the quintessential award given by the National Science Foundation to faculty performing science and engineering research," said Richard Schoephoerster, dean of UTEP's College of Engineering. "It is an award that makes a faculty member's career, and it is a mark of distinction for a faculty member. Having faculty win these awards is also a mark of a tier-one institution."
"The award (also) will impact a large number of UTEP students, giving them an opportunity to learn skills that the marketplace needs right now," Schoephoerster said. "Dr. Kiekintveld's research is at the forefront of applying gaming theory to decision-making. The results of this research will have significant impact on defense and homeland security programs."
Two of the UTEP students involved in Kiekintveld's advanced research said they probably would not get this kind of hands-on opportunity at another university so early in their college careers.
"I find it exciting to be researching in this field. It's just like the science fair projects I did in grade school only at a much higher level and in an area that's always changing," said Oscar Veliz, 24, a computer science doctoral student at UTEP. "You don't have to be a geek, but being interested in computer science and the scientific method certainly helps."
Curtis Chambers, 25, a UTEP senior majoring in computer science, said "working with Dr. Kiekintveld has been a very rewarding and awesome experience. I am fortunate that as an undergraduate student I've had the chance to be a part of this. It's also helped me to determine what I want to study for my own Ph.D. work."
"Dr. Kiekintveld's work benefits the community is various ways," Chambers also said. "His research is intuitive and innovative. He is a very, very smart man."
Kiekintveld said that recently he's worked on applications of game theory to infrastructure protection, including a decision support system called IRIS that is used by the Federal Air Marshals Service, and a second system that is in testing for nationwide deployment by the Transportation Security Administration.
"Another area that I have been heavily involved in is automated trading agents for e-commerce applications," he said. "I participated in the Trading Agent Competition, and was a lead developer for the Deep Maize agent, which won the tournament in 2008 and has consistently been among the top few performers in every tournament to date."
Kiekintveld has been at UTEP since 2010. He completed a doctoral degree in computer science and engineering at the University of Michigan, and conducted postdoctoral work at the University of Southern California.
Most Popular Stories
- Hezbollah Chief's Assassination Claimed by Sunni Group
- Allstate Seeks to Invest in Minority Firms
- SpaceX's Satellite Launch Is 'Game-Changer'
- Latin Music Conference Turns 25
- U.S. Growth Stayed Steady During Shutdown, Fed Says
- Newtown Massacre Heard on 911 Recordings
- Reid Confident Congress to Pass Immigration Bill
- Guardian Pressured to Stop NSA Stories: Editor
- New Home Sales Shoot up 25 Percent in October
- Climate Change Early Warning System Urged