Exploring the Frontier of AI and Quantum Computing

Community members joined researchers from academia and industry at Columbia University’s Faculty House on April 15 for a panel discussion exploring recent developments in two critical fields: artificial intelligence and quantum computing.

The event was a collaboration between the New York Hall of Science and Institute for Artificial and Natural Intelligence (ARNI), a new NSF-funded research institute aimed at exploring the common principles underlying natural and artificial intelligence.

Garud Iyengar, senior vice dean of research and academic programs at Columbia Engineering, welcomed the attendees.

“This institute has already become a hub for leading researchers from across the country who are committed to simultaneously better understanding the brain and building better brain-like algorithms,” he said. “We expect this event will begin a long-lasting collaboration between the New York Hall of Science — a national leader in STEM education — and ARNI as the organizations partner to create events and activities that engage the youth and families in understanding the brain its relationship to AI.”

Minerva Tantoco, the New York Hall of Science’s interim president and CEO, emphasized that her organization’s collaboration with ARNI advances its mission as a global leader in informal STEM learning.

“Today is an inspiring opportunity for us to hear about the latest advancements in these technologies and to examine how we can make these topics accessible for all learners and create a pipeline for the next generation to become future leaders and innovators in these fields,” Tantoco said. 

The conversation featured three renowned experts: Dario Gil, senior vice president and director of research at IBM, leads the company’s innovation efforts and directs research strategies in areas including AI, cloud, quantum computing, and exploratory science. Computational neuroscientist Xaq Pitkow, associate director of ARNI and associate professor of neuroscience at Carnegie Mellon University, develops mathematical theories of the brain and general principles of intelligent systems. Computer scientist Jeannette M. Wing, Columbia’s executive vice president for research and professor of computer science at Columbia Engineering, served as moderator. Before joining Columbia, Wing served as Assistant Director for the Computer and Information Science and Engineering Directorate at the National Science Foundation and led research at Microsoft.

A historic moment for computing

The panelists began by discussing what AI and quantum computing share in common. Gil, who is a member of the National Science Board, which oversees the NSF, offered an impressive historical sketch of the many diverse scientific and technological advancements that led to this point.

“We are fortunate to be living in what is arguably the most exciting time in computing since maybe the 1940s,” he said. At that time, advancements in computer science theory and the invention of the transistor allowed researchers to pursue the goal of encoding knowledge into a form computers could understand and process using formal logic. “The world saw the explosion of digital information,” he said. A few decades earlier, neuroscientist Santiago Ramón y Cajal — who, Gil noted, was a fellow Spaniard — had discovered the structure of neurons, which underlie the amazing capabilities of the human brain. These advancements were united in the late 1950s with the invention of the first neural networks. In the 1960s and ‘70s, “a third piece of the equation” fell into place with advancements in quantum physics.

“I like to summarize these vectors as bits, neurons, and qubits,” he said. Qubits (short for “quantum bits”) are the basic unit of quantum information. The vast capabilities of quantum computers stem from the fact that qubits can exist in many states while the bits in normal computers can only represent two states, 1 or 0. “A quantum computer with 100 qubits can represent information in a very rich fashion,” he said. “To represent that much information with a classical computer, I’d have to devote every atom on Earth to storing 1s and 0s.”

Speaking about AI, Pitkow made the point that theoretical advancements in 2006 and 2012 — combined with the lucky fact that computers running modern neural networks benefited from progress in designing computers for video games — have tremendously accelerated the size and capabilities of AI systems.

“Now we’re talking about putting trillions of transistors on chips and running models with trillions of parameters,” he said. “The number of neurons in your brain is less than 100 billion, and the number of connections is 100 trillion. We are now getting to the point where the scale is comparable to the scale of human brains,” Pitkow said.

“This decade, we’re going to see what happens when we combine all of these technologies to solve problems that we couldn't solve before,” Gil said.

Engineering for humanity

Gil and Pitkow agreed that the benefits of these technologies will come from how they’re deployed in real-world contexts, not from their sheer computational power.

“These are tools, and we need to use them as deliberately as we use other tools,” Pitkow said. “If you’re able to solve protein folding, it may be a struggle to use that same algorithm to solve fusion and get unlimited energy.”

Solving important problems also requires well-designed datasets and collaboration with subject matter experts. Pitkow pointed to ARNI’s partnership with NYSCI as an example of the kind of collaboration that will improve the impact that emerging technologies will have on the entire population.

“There are some huge imbalances in the datasets that we use, the problems that we're looking at, and the questions people are asking of these AI tools,” Pitkow said. “We need to have more involvement from a broader range of people who say ‘this is an important problem for us.’ To solve humanity's problems with these tools, we need to be talking with humanity.”

Gil said he was most excited about the possibility of using these breakthrough technologies to advance the process of scientific discovery. Where it could ordinarily take decades and a billion dollars to discover a more efficient industrial catalyst or improve a fertilizer, for example, these new tools could empower researchers to make such progress in just a few years and for far less money.

“The scientific method has proven to be a very effective way of answering questions to which we don’t yet have answers,” he said. “What’s exciting about AI and quantum as a combination of tools is that they could accelerate the rate of discovery.”

About NYSCI

Located in Queens — America’s most diverse county — NYSCI is committed to creating a world where diversity unlocks innovation, and where people learn to use science, technology, engineering, and math to tackle complex 21st century challenges.

NYSCI is more than a world-class destination for learning and play. It is where exhibits inspire visitors, where young scientists get their start, where community members come to learn, and where critical STEM education research occurs.

NYSCI has a strong track record of success, welcoming 500,000 visitors annually, providing professional development for more than 2,500 local teachers, and offering 2,000 high-school and college students mentoring opportunities with STEM professionals. Over the last three decades, 95 percent of the more than 4,000 young people participating in our Science Career Ladder program have gone on to college, and 70 percent of NYSCI alumni work in STEM fields.

About ARNI

The NSF AI Institute for Artificial and Natural Intelligence (ARNI) connects a decade of progress in AI to the revolution in our understanding of the brain, in order to identify common principles in human and artificial intelligence and develop a brain-like AI. Furthermore, our understanding of the brain will benefit from insights generated from AI learning algorithms and of fundamental requirements for learning derived from data. Achievements in AI will improve our understanding of how the brain processes information and accomplishes complex cognitive tasks.

ARNI meets the urgent need for new paradigms of interdisciplinary research between neuroscience and AI. ARNI pursues both theoretical and use-inspired work that will accelerate progress in both fields, and broaden their transformative impact on society in the next decade.

ARNI is a collaboration among Columbia, Baylor College of Medicine, Carnegie Mellon University, City University of New York, Harvard, Princeton, Howard Hughes Medical Institute, Mila Quebec AI Institute, Tuskegee University, the University of Pennsylvania, UTHealth Houston, and Yale. Industry partners include Amazon, Google DeepMind, IBM, and Meta, and outreach partners include the Neuromatch Academy and the New York Hall of Science. In addition to receiving NSF funding, ARNI is funded by a partnership between NSF and the Office of the Under Secretary of Defense for Intelligence and Security (R&E).