AI for Science Center

The AI for Science Center trains students to design and build AI methods that accelerate scientific discovery on their own. Coursework centers on real scientific problems - protein folding, materials design, weather and climate modeling, and mathematical reasoning - and requires students to construct and validate scientific foundation models, physics-informed and equivariant neural networks, and simulation-based inference pipelines. Students also join cross-division AI4Science projects, learning to integrate data and requirements across scientific disciplines. Graduates leave able to architect an AI pipeline for a novel scientific problem from the ground up.

⚙️ GeoAcademy System Administration College 🖥️ AI & Computing Sciences College 🔬 Basic Sciences College ⚡ Intelligent Geophysical Exploration College 🛢️ Resource & Energy Engineering College 🌍 Applied Geoscience Solutions College 💼 Economics, Policy & Strategy for the Future College 🎓 Education & Training Development College 🚀 Innovation & International Collaboration College

Machine Learning Department Computer Vision Department Natural Language Processing Department High-Performance Computing Department Quantum Computing Department Generative AI & Foundation Models Center Reinforcement Learning & Agents Center AI Safety & Alignment Center AI for Science Center Robotics & Embodied AI Center

🤖

🔑

⭐ Marvin Minsky

Chair

Researchers (real GeoVerse Lab members) 16

🔑

1901–1982

Kurt Otto Friedrichs

Researcher

💡 Co-founder of the Courant Institute; foundational contributions to the theory of partial differential equations (Friedrichs extension, symmetric hyperbolic systems, Courant-Friedrichs-Lewy condition) underpinning numerical PDE solvers

🔑

1906–1998

Jean Leray

Researcher

💡 Founder of algebraic topology methods (Leray-Schauder degree theory, spectral sequences) used to prove existence of PDE solutions; his fixed-point/degree theory underlies convergence analysis of operator-learning networks

🔑

1914–2002

Max Perutz

Researcher

💡 Solved the first three-dimensional structure of a protein (hemoglobin) by X-ray crystallography, founding structural molecular biology and the structure-function paradigm that AlphaFold-class models now automate

🔑

1901–1994

Linus Pauling

Researcher

💡 Established the quantum-mechanical theory of the chemical bond and empirical rules for predicting crystal structures (Pauling's rules), the foundation of rational and now AI-driven materials design

🔑

1881–1953

Lewis Fry Richardson

Researcher

💡 Pioneered numerical weather prediction by manually solving the primitive equations for a forecast (1922), founding the field that GraphCast-class AI weather models now accelerate by orders of magnitude

🔑

1918–2013

Frederick Sanger

Researcher

💡 Developed the sequencing methods (protein sequencing and DNA dideoxy sequencing) that generated the sequence data underlying all of modern genomics and the deep-learning genomic models trained on it

🔑

1925–2004

John Pople

Researcher

💡 Developed Gaussian-basis-set ab initio computational quantum chemistry methods (and the Gaussian software) that generate the labeled quantum-mechanical data used to train modern molecular ML potentials

🔑

1881–1963

Theodore von Kármán

Researcher

💡 Founded modern aerodynamics and turbulence theory (Kármán vortex street, Kármán momentum-integral method), the classical fluid-dynamics benchmarks that neural PDE surrogates for turbulent flow are validated against

🔑

1930–2016

John Alan Robinson

Researcher

💡 Invented the resolution principle and unification algorithm (1965), the foundational inference mechanism of automated theorem proving and logic programming that underlies today's AI-assisted mathematical discovery systems

🔑

1898–1974

Fritz Zwicky

Researcher

💡 Pioneered large-scale systematic sky surveys (Palomar Sky Survey) and inferred dark matter from galaxy cluster dynamics, founding the data-driven, survey-scale astronomy that AI classification and discovery pipelines now operate on

🔑

1918–1999

Gertrude B. Elion

Researcher

💡 Pioneered rational drug design by targeting differences in biochemical pathways between pathogens/cancer cells and normal human cells, founding the target-based design paradigm that generative AI drug discovery now automates

🔑

1919–2013

George E. P. Box

Researcher

💡 Founded modern statistical model-building philosophy ('all models are wrong, but some are useful') and response-surface/design-of-experiments methodology, foundational to simulation-based inference and surrogate-model uncertainty quantification

🔑

1931–2019

Wallace Smith Broecker

Researcher

💡 Founded quantitative Earth system science, coining 'global warming' (1975) and formulating the ocean 'conveyor belt' thermohaline circulation model that Earth-system AI emulators now build on

🧑‍🔬

🔑

1898–1969

Warren Sturgis McCulloch

Researcher

💡 Co-created (with Walter Pitts) the first mathematical model of a neuron as a logical/computational unit, founding both computational neuroscience and the artificial neural network on which all modern deep learning descends

🔑

1894–1981

Jerzy Neyman

Researcher

💡 Founded the potential-outcomes framework for causal inference and modern hypothesis-testing theory (Neyman-Pearson), the statistical foundation underlying today's AI-assisted causal discovery methods

🔑

1910–1994

Dorothy Hodgkin

Researcher

💡 Determined the structures of penicillin, vitamin B12, and insulin by X-ray crystallography, generating the gold-standard structural datasets and phasing methodology that structure-prediction AI is trained and validated against