BEGIN:VCALENDAR
VERSION:2.0
CALSCALE:GREGORIAN
PRODID:UW-Madison-Physics-Events
BEGIN:VEVENT
SEQUENCE:2
UID:UW-Physics-Event-9810
DTSTART:20260827T143000Z
DTEND:20260827T163000Z
DTSTAMP:20260827T070010Z
LAST-MODIFIED:20260824T165829Z
LOCATION:B343 Sterling
SUMMARY:Physics and Machine Learning: How Each Can Benefit from the Ot
 her\, Preliminary Exam\, Raheem Hashmani\, Physics PhD Graduate Studen
 t
DESCRIPTION:The intersection of physics and machine learning (ML) allo
 ws both fields to benefit from each other. By treating ML as a physica
 l system\, we study it as a phenomenon to be understood through contro
 lled experiments\, well-defined datasets\, and falsifiable hypotheses.
  Conversely\, by using ML as a tool\, we can accelerate discoveries in
  physics across a variety of fields. This talk covers 3 projects spann
 ing these topics. A large portion of the talk is dedicated to our work
  on developing physics-style methods using exact likelihoods\, inducti
 ve biases\, and information theory to probe what these models are real
 ly learning and explore the edge cases where they succeed or fail. Our
  first project models two hypotheses as Gaussian processes\, allowing 
 us to generate data for which we can compute the exact likelihood of a
  sample under both hypotheses. This allows the construction of a Bayes
  optimal classifier\, an architecture-independent upper bound on perfo
 rmance\, and the benchmarking of real classifiers against it to isolat
 e the role of inductive bias. Our second project introduces a framewor
 k for generating highly multimodal datasets with explicitly calculable
  mutual information (MI) between modalities. This enables controlled e
 xperiments that systematically probe how mutual information and its di
 stribution across modalities affect multimodal self-supervised learnin
 g and mutual information estimation. The remaining portion of this tal
 k focuses on facilitating the use of ML in the designing of fusion dev
 ices. Our third project introduces Driftless Star\, an open-source wor
 kflow that computes transport-consistent plasma profiles for a given s
 tellarator boundary and plasma conditions in a unified\, reproducible 
 pipeline. The pipeline enables AI-accelerated fusion device design by 
 allowing neural surrogates to replace costly physics stages and large 
 language model agents to orchestrate the chain\, make judgment calls b
 etween stages\, and autonomously search for improved geometries.
URL:https://www.physics.wisc.edu/events/?id=9810
END:VEVENT
END:VCALENDAR