What makes something intelligent? Where is intelligence to be found? How is intelligence studied? SFI researchers Melanie Mitchell and Melanie Moses have organized a virtual conference in March that aims to answer questions about the foundations of intelligence from areas as diverse as philosophy to evolutionary intelligence and complex information processing.
“We're trying to get a sense of where people are when thinking about what intelligence is, what's important to study, what are the big open questions,” says Mitchell, who is SFI’s Davis Professor of Complexity.
Mitchell cites AI pioneer Marvin Minsky, who referred to intelligence as a “suitcase word”—one filled with all kinds of definitions. There’s human intelligence and the artificial intelligence of robots, but there’s also intelligence in swarms of ants and maybe even a kind of intelligence in something like a market. There’s evolutionary intelligence as well as social intelligence.
For three hours every morning during the week of March 15, presenters at the meeting will discuss these topics, trying to puzzle together just what intelligence means. Each presenter will speak for 30 minutes, and then there will be 20 minutes for discussion and questions, to facilitate the kinds of conversations that would usually happen outside of talks at an in-person conference. The list of confirmed speakers includes SFI External Professors, resident Faculty, Postdocs, and researchers from beyond SFI, such as Daniel Dennett, Ricard Sole, Stephanie Forrest, and Alison Gopnik, whose fields of study span philosophy, physics, biology, computer science, and psychology.
“We're inviting basically everyone in the SFI community,” says Mitchell. She hopes the meeting will get people out of their narrow research areas and think about the big picture.
The goal of this meeting is not to settle the matter and find answers to all of these questions about intelligence, but to forge interdisciplinary collaborations. The meeting is supported by a National Science Foundation “planning” grant, with the aim to develop a full proposal for the NSF’s AI Institute program.
“We hope to generate ideas in these workshops that will inform the full proposal writing,” says Moses, an external professor at SFI. “We hope that a subset of the participants will also be co-PIs or personnel on the full proposal.”
Five other meetings are planned for 2021-2022, and depending on COVID conditions, some may be in person.