Search
  • Artificial Intelligence
  • News
  • Artificial Intelligence

Playing the long game with AI

by Krista Davidson Nov 12 / 20

Canada CIFAR AI Chair Kevin Leyton-Brown applies AI to microeconomics to predict the best possible outcome based on our interests. 

2014-06-Midway
Photo Credit: Paul Joseph for UBC Brand and Marketing

His area of expertise in AI is game theory, a computational approach to understanding how different players or agents with competing interests interact with each other, and how those interactions shape different outcomes. 

More than just a computational approach, game theory guides the decisions of businesses, governments and individuals. It predicts how the public will respond to new products, how voters will respond to policies, and provides insight into stock market investments and mortgage rates. Game theory has become increasingly important to the private and public sector. Leyton-Brown wants to use it for change.

He is passionate about using AI technologies to solve important challenges to humanity.

“AI is largely funded by industry so it tends to serve a narrow set of industrial needs. There is more research on using AI to make effective advertisements than there is in identifying vaccine candidates for rare diseases,” says Leyton-Brown.

“There are many corners of the world that go untouched because they don’t match up with corporate interests, but those are the areas where AI could really move the needle,” he explains. 

Leyton-Brown lived in Uganda for a part of a 2010 sabbatical.  During this time he was particularly moved by a situation that could be improved with AI. He noticed that vendors who lived in the city were able to charge competitive prices for their produce, and often sold out before the end of the day. Meanwhile, their suppliers in more remote areas struggled to find a market for their produce even at low prices. Given that most Ugandans have cell phones, he developed and co-founded Kudu, a mobile app that enables farmers to locate and trade with vendors for competitive prices.

Another project involved a collaboration with the U.S. Federal Communications Commission (FCC). His team developed an algorithm that enabled the FCC to package radio spectrums in an auction. The auction was so profitable that it yielded the government with over $7 billion to pay down the national debt. The availability of frequencies enabled the cellphone industry to expand the market and provide the public with better cell service. 

The Natural Sciences and Engineering Research Council awarded him the 2014 NSERC E.W.R. Steacie Memorial Fellowship for his research, noting that his  work on the auction of broadcast airwaves would likely be replicated throughout the world. He also received the Franz Edelman Award for Achievement in Advanced Analytics, Operations Research and Management Science.

“We took something that didn’t make sense anymore and reallocated it for a positive impact all around,” he says, referring to the project of repurposing radio spectrum from broadcast TV to mobile broadband.

In addition to his research and consulting for companies, he also teaches a course about AI for social impact to graduate students at the University of British Columbia. In a recent course, students developed a system using commercial satellite images of fallow grasslands to predict soil moisture. Using existing climate models and crop development models, the students showed how their system can provide irrigation recommendations to increase crop yields in low-income countries.

Leyton-Brown is an Amii fellow, a professor at the University of British Columbia’s department of computer science, and director of the UBC ICICS Centre for Artificial Intelligence Decision-making and Action. He completed his graduate studies at Stanford University, collaborating with Paul Milgrom, a recipient of the 2020 Nobel Prize for Economic Sciences.