By Ian Scheffler
In 1943, Warren McCullough, a psychiatrist at the University of Illinois at Chicago, and Walter Pitts, a runaway prodigy, co-authored a paper in the Bulletin of Mathematical Biophysics on the relationship between biological neurons and formal logic.
As Brian Christian recounts in “The Alignment Problem,” the idea that the logic at the heart of computers could model neurons was decades ahead of its time. Thanks to the exponentially higher amount of computing power available today, what McCullough and Pitts proposed — using logic to model the activity of neurons — has become the foundation of artificial intelligence today, powering everything from image recognition systems to ChatGPT.
The story of AI at Penn Engineering has likewise been decades in the making. Just as generations of scientists in diverse fields contributed to the ideas now driving AI forward, a range of philanthropically minded Penn Engineers and friends of Penn Engineering have underwritten the programs, physical structures and research initiatives that make the School a leader in training the next generation of engineers in AI.
That story begins in 2003, with the creation of “Networked Life,” a course taught by Michael Kearns, National Center Professor of Management & Technology in Computer and Information Science. For the first time, that course brought together disciplines from inside and outside engineering that relate to networks. “We haven’t asked engineering students to take a course in game theory to understand how incentives work or in sociology to understand human behavior,” Kearns told the Penn Almanac at the time. “There is now enough science out there on the intersection of these topics to design undergraduate courses.”
In 2003, social networks were just beginning to take off. MySpace launched the same year; Facebook was a year away. Enrollment in the course exploded, showing its popularity, which is how Rajendra Singh (PAR’10, PAR’11), a member of the Penn Engineering Board of Advisors, and his wife Neera (PAR’10, PAR’11), both pioneers in developing cellular networks, heard about the class a few years later. In 2009, the Singhs partnered with Penn Engineering to announce a new major, expanding the interdisciplinary focus of Networked Life into what is now known as the Networked and Social Systems Engineering (NETS) program. For the first time, students could combine the study of algorithms, game theory and more in a complete degree, not just a course, to study networks themselves.
It wasn’t long before Fred Warren (ME’60, WG’61, PAR’87, PAR’19) another Board member and perhaps Penn Engineering’s first dual degree graduate, took notice. In 1956, his first year at The Moore School (as part of Penn Engineering was then known), Warren approached S. Reid Warren (no relation), the then-Associate Dean of Undergraduate Engineering Affairs, about combining studies in business and engineering. “He said, ‘Get your hat and coat and let’s go down the street to the Wharton School.’” That special arrangement led Warren to receive degrees from both Penn Engineering and Wharton, foreshadowing Penn’s longest-running coordinated dual-degree program, the Jerome Fisher Program in Management & Technology.
Warren saw the benefits of interdisciplinary learning firsthand during his career in venture capital. His training in both engineering and business helped him make prescient decisions like believing a twenty-something Steve Jobs, at a meeting with investors in 1978, when Jobs predicted that everyone would one day own a computer.
In 2013, just two years after NETS enrolled its first cohort, Warren and his wife Robin announced the creation of the Warren Center for Network & Data Sciences to support interdisciplinary graduate and faculty research related to networks. “It was really a takeoff on what Raj was doing at the undergraduate level,” says Warren.
Shortly thereafter, Harlan Stone (C’80, PAR’13) a University Trustee who also sits on the Penn Engineering Advisory Board, found himself drawn to the Warren Center’s interdisciplinary focus. “I saw programs like the Warren Center,” Stone says, “where suddenly you’re crossing out of science and engineering into social science, which is closer to the interests in my heart.”
A leader in the design and manufacture of sustainable decorative building materials, Stone has always straddled the arts and sciences. After learning about the Warren Center and Penn Engineering’s undergraduate program in Digital Media Design, which partners with the Annenberg School for Communication and the Weitzman School of Design, he began to hatch an idea for a new project to benefit the University. “I thought maybe the intersection of the humanities and data science could be something like a new renaissance,” Stone says.
Initially, his conversations with Eduardo Glandt, then-Dean of Penn Engineering, sketched out a center studying visual perception. “I started thinking more and more about the impact,” says Stone. “I kept thinking, it’s not about data science helping the humanities. I got this moment where the lightbulb went off, where it was the humanities and the liberal arts helping big data and engineering to imagine a world in which we use this incredibly empowering technology to do good.”
In 2017, Penn Engineering announced plans for Amy Gutmann Hall, a new home for data science across disciplines and schools at Penn, which will open later this year. Located near the heart of campus, Amy Gutmann Hall will be within walking distance of every school at Penn. “The physical intersection is 34th and Chestnut,” says Stone, who provided the building’s naming gift. “But it’s also the intersection of big data, and thoughtfulness, and humanity, and ethics, and caring, and empathy, and hopefully compassion — that’s the real intersection.”
Months later, a handful of researchers at Google published “Attention is All You Need,” the paper underlying many of today’s most prominent AI tools. The paper proposed a novel architecture for training neural networks, the descendants of McCullough and Pitts’ ideas.
“When ChatGPT and all the other systems started coming out,” says Rajendra Singh, “it became quite clear to me that, as far as artificial intelligence was concerned, it’s not going to be limited to a certain set of specific applications. It’s going to have a more general use and can be applied to almost anything.”
Seeing these rapid technological changes, the Singhs couldn’t help but think of their early days in cellular communications. “When the wireless industry started,” Singh recalls, “nobody thought that more than 2% of the population would have cell phones.” Today, according to the World Economic Forum, cell phones outnumber people.
Earlier this year, Penn Engineering announced the creation of the Raj and Neera Singh Program in Artificial Intelligence, which will include the first Ivy League Bachelor of Science in Engineering in AI and the first online Ivy League Master of Science in Engineering in AI. The undergraduate program will be housed in Amy Gutmann Hall and the online master’s program will expand access to cutting-edge AI education.
Both programs will combine resources from the Departments of Computer and Information Science (CIS) and Electrical and Systems Engineering (ESE) to prepare students for careers in AI. “I think education in any new idea or technology is the most wonderful thing,” says Singh, “because these are the people that go out and spread that knowledge to the entire society.”
To the Singhs, the moral component of the new programs, which will include courses like “Ethical Algorithm Design” taught by Michael Kearns, is just as important as the technical content. “Any technology can be used for good or bad,” says Singh. “You hope that in a setting like the University of Pennsylvania, where you have people who think about the world in general, about ethics and law, they will steer it more in the direction of being good for mankind.”
Watch the full conversation with Dean Kumar, Rajendra Singh, Harlan Stone and Fred Warren