Course positioning
EE 292R (CS 292/PSYCH 292R/STATS 282) is a 1-unit seminar first offered in Autumn 2026, organized by Candès, Wetzstein, Narasimhan, and Wandell. It does not teach specific techniques; each week a speaker presents a real AI4Science case, covering neuroscience, medicine, genomics, computational social science, world models, and more.
What it covers
- Neuroscience: retina/visual system, brain foundation models, digital twins
- Medicine: medical imaging FMs, precision medicine, multimodal omics
- Genomics: regulatory DNA, non-coding variants
- Spatial intelligence: video world models, world agents
- Computational social science: coordinated behavior detection across encrypted platforms
- Research infrastructure: Marlowe GPU cluster, petabyte-scale data
Prerequisites
No hard prerequisites; interest in AI + science is enough.
Assessment & workload
Attendance plus a short write-up on one lecture topic; 1 unit, S/NC.
Who it's for
People who want a fast map of AI4Science and to find interdisciplinary topics/collaborators; not for people who want systematic ML technique training.
Study advice
Lectures are in person only, but Stanford Online will upload recordings to YouTube. Write down one 'method + scientific question' pair per talk.
Related courses
Official course description
AI4Science is an emerging discipline that leverages machine learning, foundation models, and autonomous agents to accelerate the pace of discovery. By integrating data-driven architectures with physical constraints, the field models complex systems at unprecedented scales. At Stanford, this work is empowered by Marlowe - a specialized GPU-based computational instrument. This seminar series presents research from the Stanford community and beyond, exploring the frontier where Al and scientific inquiry converge.