Skip to main content

Matter of Minds

Building the Science of Natural and Artificial Intelligence

Nov. 4, 9:30am–3:30pm
logo for Matter of Minds

Learn about recent research in natural intelligence and current topics in AI at MIT.

On Monday, November 4, 2024, the Quest hosted a day of research and demonstration from the neuroscientists, cognitive scientists, computer scientists, and software engineers who are creating the science of intelligence. 

MIT's Quest for Intelligence couples scientific enquiry and rigorous engineering to address real-world challenges that are beyond machine capabilities but within the ability of natural intelligence.

Recordings of sessions are available.

Speakers

Schedule

Morning session

9:00 AM - 9:45 AM

Registration

9:45 AM - 10:00 AM

Welcome Remarks MIT President Sally Kornbluth

10:00 AM - 10:30 AM

What the Quest Is / A Roadmap for the Day Professor Jim DiCarlo, Director of the MIT Quest for Intelligence

10:30 AM - 11:45 AM

Language and the Development of Thought: A conversation led by Professor Josh Tenenbaum, BCS

Rebecca Saxe, John W. Jarve Professor of Cognitive Neuroscience (BCS), Associate Dean of the School of Science

Laura Schulz, John and Dorothy Wilson Professor and Associate Department Head (BCS)

Ev Fedorenko, Associate Professor (BCS)

Jacob Andreas, Associate Professor (EECS)

Exhibitions and Demonstrations (Lunch)

12:00 PM - 1:00 PM

Spatiotemporal Pattern Learning The ability to learn a wide variety of structured concepts from sparse, ambiguous input is one of the hallmarks of human intelligence. This study probes the evolutionary and developmental origins of this ability by comparing the performance of human adults, children, and macaque monkeys in a simple but surprisingly rich and open-ended spatiotemporal sequence prediction task: guessing "Where will the dot go next?"

VR: Find the Orange Interact with a VR tabletop to locate and grasp items.

Exploring Human Visual Perception with High-End VR: Insights from Real and Altered Worlds Immerse yourself in realistic and altered environments, using the Quest’s new VR infrastructure. This demonstration showcases the new VR infrastructure at the MIT Quest, designed to facilitate cutting-edge research on visual perception, featuring the Varjo XR-4 professional headset.

Word to World Models: Induction & Deduction in the Human Brain The capacity to induce novel logical structures from experience and deduce what follows from them is ubiquitous in humans. This project explores the neural basis of inductive and deductive logical reasoning — particularly in the context of our brain’s other high-level processing networks for language and for abstract domain-general cognition.

Brain-Score: Which Artificial Neural Network is the most Brain-Like? Brain-Score is a comprehensive benchmarking platform that evaluates how well computational models match neural and behavioral data from the brain.

PsychoScope: A scalable measurement tool for human cognition Crowdsourced, large-scale datasets of behavior are a key component in building modern AI systems – and increasingly, in building and verifying scientific models of cognition. PsychoScope is a software platform which enables non-experts to seamlessly collect, share, and compute on high-quality, gold-standard empirical measurements of human cognition.

Stimulus Library/iCatcher+ Stimulus Library, supported by the MIT Quest for Intelligence and Lookit, is an open registry of reusable experimental stimuli. By enabling the sharing and reuse of stimuli, Stimulus Library reduces the time required to set up experiments and fosters collaboration among researchers. iCatcher+ is a tool to assist researchers working with children.

GenBrain Cortico-cortical computation as real-time spiking neural inference in probabilistic programs

ChiSight real-time 3D object perception via probabilistic programming

GenJAX an open-source platform for massively parallel probabilistic programming

Afternoon session

1:15 PM - 2:30 PM

Building Machines and Making Them Intelligent A conversation led by Professor Nick Roy, Aero-Astro

Leslie Kaelbling, Panasonic Professor (EECS), Quest Director of Research

Nancy Kanwisher, Walter A. Rosenblith Professor of Cognitive Neuroscience (BCS)

Jim DiCarlo, Peter de Florez Professor (BCS), Director of the Quest for Intelligence

Vikash Mansinghka, Principal Research Scientist (BCS), Quest Director of Modeling and Inference

2:30 PM - 3:00 PM

Fireside Chat with Dr. David Siegel SM ’86 PhD ’91, hosted by Dean Dan Huttenlocher, Schwarzman College of Computing

3:00 PM - 3:15 PM

Closing Remarks Professor Jim DiCarlo