Minds, Brains, and Machines (MBM)

Overview

Understanding intelligence is one of the greatest scientific challenges, demanding an interdisciplinary approach that spans psychology, neural science, philosophy, linguistics, data science, and artificial intelligence (AI). At the core of this endeavor is computation—viewing intelligence as an advanced, adaptive computational process. However, the exact nature of the computation required for intelligence remains an open question. Despite striking recent progress in AI, current technologies provide nothing like the general-purpose, flexible intelligence inherent in humans.

NYU Minds, Brains, and Machines is a campus-wide initiative launched by FAS Dean Antonio Merlo. It aims to foster research and educational opportunities around these topics. The initiative started with a conference on human and machine intelligence and now prioritizes a hiring cluster that connects several academic units at NYU, including CDS, Psychology, the Center for Neural Science, and the Flatiron Institute.

Colloquium Series

The colloquium series showcases some of the most exciting work at the intersection of human and machine intelligence. Lectures will be held in person when possible.

Fall 2024 Events

David Bau (Northeastern University) and Grace Lindsay (CDS) – joint event with CDS Seminar Series

Date and Time: November 1, 2024, 2:00pm

Abstract: In this talk we discuss recent work in interpreting and understanding the explicit structure of learned computations within large deep network models. We examine the localization of factual knowledge within transformer LMs, and discuss how these insights can be used to edit behavior of LLMs and multimodal diffusion models. Then we discuss recent findings on the structure of computations underlying in-context learning, and how these lead to insights about the representation and composition of functions within LLMs. Finally, time permitting, we discuss the technical challenges of doing interpretability research in a world where the most powerful models are only available via API, and we describe a National Deep Inference Fabric that will offer a transparent API standard that enables transparent scientific research on large-scale AI.

Scientific Bio: David Bau is an assistant professor at Northeastern University Khoury school of Computer Sciences. He is a pioneer on deep network interpretability and model editing methods for large-scale AI such as large language models and image synthesis diffusion models. He is leading an effort to create a National Deep Inference Fabric.

Inbal Arnon (Hebrew University of Jerusalem)

Date and Time: September 23, 2024, 11:00am

Abstract: The learnability consequence and sources of Zipfian distributions: how cultural evolution can give rise to distributional properties of language. While the world’s languages differ in many respects, they share certain commonalities: these can provide crucial insight into our shared cognition and how it impacts language structure. In this talk, I explore the learnability consequences and sources of one of the most striking commonalities across languages: the way word frequencies are distributed. Across languages, words follow a Zipfian distribution, showing a power law relation between a word’s frequency and its rank. The source of this distribution has been heavily debated with ongoing controversy about what it can tell us about language. Here, I propose that such distributions confer a learnability advantage, leading to enhanced language learning in children, and the creation of a cognitive pressure to maintain similarly skewed distributions over time. In the first part of the talk, I will examine the learnability consequences of Zipfian distributions, focusing on their greater unigram predictability. In the second part, I explore the learnability sources of Zipfian distributions to ask whether learning biases can help explain why such distributions are so common in language. I will present joint work with Prof. Simon Kirby suggesting that this foundational distributional property of language can emerge through cultural transmission characterised by whole-to-part learning. I will end by discussing implications of these findings for large language models.

Scientific Bio: Prof. Arnon has a PhD in Linguistics and Cognitive Science (2011, Stanford University), and is currently a Full Professor of Psychology at the Hebrew University. Her research program, which lies on the interaction of Linguistics, Psychology, and Cognitive Science, focuses on understanding human’s unique ability to learn, use, and develop language, and more specifically, on understanding how children acquire language, how they differ from adults in doing so, and how learnability pressures shape the emergence and structure of human language. Prof. Arnon has worked extensively on first language acquisition, developing a novel framework for understanding why children are better language learners than adults, with applied implications for human and machine learning (The Starting Big Approach, see Arnon, 2021 for a review). In her current projects, she asks whether learning pressures and constraints can explain why languages look the way they do, and how linguistic structure emerged to begin with.

Job Opportunities

There are no current open faculty searches but stay tuned!

Key NYU Faculty

Robert J. Glushko Undergraduate Thesis Prize

The Glushko Prize for Outstanding Undergraduate Honors Thesis in Minds, Brains, and Machines is awarded annually to an NYU student who has conducted an Honors thesis in computational cognitive science or otherwise at the intersection of human and machine intelligence. Students in the Department of Psychology are eligible to apply if they complete an honors thesis or other independent thesis project.

The winning thesis will be the one demonstrating the greatest academic excellence and creativity using computational methods to inform aspects of intelligent behavior.

Students in Center for Data Science, Center for Neural Science, Linguistics, or Psychology who would like their thesis considered for this prize should have their faculty sponsor notify Todd Gureckis (todd.gureckis@nyu.edu) by April 1 of the spring semester (theses completed in the previous fall are still eligible for the spring).