Filling the Multi-Agent Void

Abstract: When AI agents meet and interact, their behavior is currently underspecified by training. There is a window of opportunity right now to seed this "multi-agent void" with examples of AI agents collaborating, with humans and with each other, in ways that humans endorse. These early examples of agent interactions can influence the training prior for the next generation, in the same way that early examples of the "helpful assistant" persona shaped the prior for today's chatbots. I'll outline a mathematical model of self-reinforcing priors, and describe a small ongoing experiment.

Speaker's bio: Lionel Levine is a professor of mathematics at Cornell University. He pivoted from pure math (probability and statistical physics) to AI safety in 2022. He advises Cornell AI Alignment and the Math+AI Lab. His research is aimed at increasing the probability that future AI systems benefit all of humanity and the planet.

Moderator's bio: Irfan Alam is an autistic mathematician, currently working as a postdoctoral fellow in the Department of Computer and Mathematical Sciences at University of Toronto Scarborough. He completed the BMath and MMath programs at the Indian Statistical Institute (2009-2014), before joining Louisiana State University as a graduate student, where he obtained his PhD in mathematics as well as MA in Philosophy in 2021. Irfan’s mathematical focus is on applications of nonstandard analysis to standard mathematics, especially in probability theory, topology, and lately in the foundations of machine learning. He also works on the philosophy of mathematical practice, with particular interests in the history of infinitesimals and in epistemic injustices in mathematics caused by various socio-cultural norms.

Computer scienceMathematics

Audience: researchers in the topic

---

Relatorium seminar

Series comments: The name "Relatorium" combines "relator" with the Latin root "-ium," meaning "a place for activities" (as in "auditorium" or "gymnasium"). This seminar series is a platform to relate ideas, interact with math, and connect with each other.

In this series, we explore math beyond what we usually hear in standard talks. These sessions fall somewhere between a technical talk and a podcast: moderately formal, yet conversational. The philosophy behind the series is that math is best learned by active participation rather than passive listening. Our aim is to “engage and involve,” inviting everyone to think actively with the speaker. The concepts are accessible, exploratory, and intended to spark questions and discussions.

The idea of relatability has strong ties to compassion — creating space for shared understanding and exploration - which is the spirit of this seminar! This is a pilot project, so we’re here to improvise, learn, and evolve as we go!

---