CAN AI MODELS BE LIKE SCIENTIFIC INSTRUMENTS?

#STEM #measurement
Share

Can AI models be like scientific instruments?

Modern ML/AI models are incredibly complex: they are expensive to train and use and often lack interpretability. In this talk, I will explore some recent results centered on the relationship between AI models and classical measurement technologies like scientific instruments. The latter are grounded in physics whereas the former are “data driven”: this leads to some very basic questions that we need to understand: how should we measure how models differ from each other and how do they change through training? We can use some tools from "old school" statistics/probability to get some handle on this in terms of understanding variability during training and using AI models as "instruments" to look at other models. While much of this work is empirical, the findings point to some interesting directions for theory and engineering.

This talk is based on joint work with Sinjini Banerjee, Reilly Cannon, Sutenay Choudhury, Tony Chiang, Ioana Dumitriu, Andrew Engel, Natalie Frank, Tim Marrinan, Max Vargas, and Zhichao Wang.


  Date and Time

  Location

  Hosts

  Registration


  • Add_To_Calendar_icon Add Event to Calendar
  • Posgrado de Ingeniería Eléctrica, Ciuda Universitaria
  • Morelia, Michoacan de Ocampo
  • Mexico 58030
  • Building: Omega II
  • Click here for Map
  • Contact Event Hosts
  • Co-sponsored by Univseridad Michoacana de San Nicolás de Hidalgo
  • Starts 04 October 2025 06:00 AM UTC
  • Ends 22 October 2025 06:00 AM UTC
  • No Admission Charge




Agenda

  • 1:00-1:105 Welcome and Distinghisehd Lecture Speaker Introduction

  • 1:10 - 2:10 Can AI models be like scientific instruments?

  • 2:10 - 2:40 Question and Answer session