Carnegie Mellon University has a long history in artificial intelligence — in fact, the longest in the world.
Long before CMU’s AI researchers were lured away by Uber to lead autonomous vehicle efforts at its Advanced Technologies Group in the Strip District, and even before U.K.-based Delphi Automotive PLC acquired the university’s spinoff, Ottomatika, for much the same reason in 2015, there was Allen Newell and Herbert Simon, researchers at what was then known as Carnegie Tech. The two men were responsible for the birth of AI in 1956.
Sixty-one years and countless innovations later, the university has developed an initiative, CMU AI, that will unite the university’s various disciplines of artificial intelligence research under one umbrella and potentially draw more startups, investors and funding to the university and region.
Artificial intelligence is a huge field of study, impacting the economy in both exciting and frightening ways at once, said Andrew Moore, dean of CMU’s School of Computer Science. That prospect means countless applications for AI, which in turn require specialization.
“There are multiple parts to [AI]. One of the strengths of Carnegie Mellon is that we are experts in all of those parts,” he said. “As we have grown, we have focused on those parts, but we have not focused enough on how those parts fit together.”
That departmental splintering made sense before, as a huge pool of researchers, professors and students worked on certain components of artificial intelligence.
The most basic level of CMU’s AI research governs sensing and understanding the world. “Think of it like an organism,” Mr. Moore said.
First, a computer must feel and understand, collecting input from the world through vision and speech. Those duties fall under the Language Technologies Institute and CMU’s Computer Vision Group.
Higher up the ladder, CMU’s Machine Learning Department — the largest of its kind in the world — teaches computers how to learn from experience.
In addition, CMU has a group devoted to AI-aided search, which is essentially a form of information retrieval. This form of AI search is not pigeonholed into use for search engines, but can be applied to robots who must retrieve information at the site of a disaster.
At CMU, AI search functions are used to retrieve data on traffic lights within one mile of the university’s campus in Oakland, optimizing traffic lights to negotiate the best system for cars funneling through intersections with the lowest number of emissions as possible.
“On top of the AI stack of tech is robotics, or the ability to act,” Mr. Moore said. “There is very little point in having a computer that learns and thinks without acting.”
These divisions of artificial intelligence are pieces of a larger whole, he said, and make more sense as an interdisciplinary department, free of hierarchical input, to create full systems, rather than individual parts.
Mr. Moore said there are 120 faculty members working in various components of AI research, among 200 other Ph.D. students, 300-400 graduate students and at least 150 undergraduates.
“It’s a very large mass of expertise,” he said.
That conglomerate of AI experts has churned out high-profile intellectual property that has been acquired by several of the Big 5 tech companies in the country.
In 2009, Google acquired reCaptcha, a form of open-source tech that uses CAPTCHAs to create a series of security questions to prevent spam and fraud. That invention was created by a CMU computer science professor, Luis Von Ahn, who also is the founder and CEO of East Liberty-based language learning startup, Duolingo.
In 2016, Facebook acquired CMU spinoff Faciometrics, a facial recognition feature founded by Fernando De la Torre, an associate research professor at CMU.
Mr. Moore said that last year marked the first time that the majority of its computer science graduates wound up in Pittsburgh, not California.
“We are getting into this economic world, now, where we have a very reasonable shot in becoming the world’s center for artificial intelligence,” he said. “Today’s initiative is to take advantage of all aspects of AI tech.”
Mr. Moore also believes that having one umbrella that encompasses all aspects of artificial intelligence will help the university gain more funding for research.
“Yes, I do think that is a big part of what’s going on here,” he said. “In the broader world of industry, medicine and defense, we have to make sure all potential partners are really aware Pittsburgh and CMU is a one-stop shop for crazy, advanced research in AI.”
Courtney Linder: clinder@post-gazette.com or 412-263-1707.
First Published: June 27, 2017, 4:48 p.m.
