NERSC Generative AI for Science Program

2024 GenAI Call for Proposals

In early 2024, NERSC invited proposals for projects that will leverage Perlmutter to push the state of the art in Generative AI (GenAI) and deep learning for science, as well as produce novel science outcomes. The call was focused on teams with expertise in the use of deep learning for science, a thorough understanding of the scientific domain, and demonstrated proofs of concept. With 55 proposals, 36 teams were awarded a total of O(300K) GPU node hours based on project scope, relevance to the NERSC mission areas, and readiness to use large computational resources at scale to advance their scientific outcome. Application areas include the use of novel GenAI models and techniques in material sciences, biological sciences, high-energy physics, weather (and climate) sciences, and applied mathematics.

The list of accepted teams (in no particular order) are:

Project PI	PI Institution	Project Title
Tong Geng	University of Rochester	GenAI-based Prediction of Hot Electron and Neutron Generation for Nuclear Fusion
Vyacheslav Bryantsev	Oak Ridge National Laboratory	Reactive Chemistry in Molten Salts
Christopher Bretherton	Allen Institute for Artificial Intelligence	Efficient generative regional precipitation downscaling for a climate model emulator
Victor Batista	Yale University	High-Throughput Screening Drug Discovery Through Generative Model and Molecular Docking
Feng Yu	Lawrence Berkeley National Laboratory	AlphaSAXS - Reconstruct Protein Ensemble from High-throughput X-ray Scattering
Massimiliano Lupo Pasini	Oak Ridge National Laboratory	Generative graph-based AI for scalable, accurate, and transferable exploration of large chemical spaces for accelerated alloy design
Nesar Ramachandra	Argonne National Laboratory	Foundation Model Ecosystem for Computational Science
Zhong Wang	Lawrence Berkeley National Laboratory	A Biosynthetic Gene Cluster Foundation Model (bgcFM) for Systematically Exploring the Secondary Metabolic Capacity Encoded by Diverse Microbes
Sha Feng	Pacific Northwest National Laboratory	Building generative AI architectures to predict local wind energy generations using larger-scale environment
Gabriele Corso	MIT	Deciphering biomolecular structures with hierarchical latent diffusion models
Gebrand Ceder	Lawrence Berkeley National Laboratory	Inverse design of amorphous and crystalline materials for superionic conductors
Omar Yaghi	University of California Berkeley	Designing novel Metal-Organic Frameworks for Carbon Capture and Water Harvesting with Diffusion Model
Jackie Zhi Yao	Lawrence Berkeley National Laboratory	Multimodal Learning Framework Using Large Language Models for the Discovery of Inorganic Materials
Shaowu Pan	Rensselaer Polytechnic Institute	Scalable Scientific Machine Learning for Complex Physical Systems
Deyu Lu	Brookhaven National Laboratory	Atomic scale simulation of aqueous zinc electrolytes using deep neural network potentials and generative diffusion models
Bonnie Berger	MIT	An Ensemble for Every Protein: Diversifying the Protein Data Bank with AlphaFold and Gen AI
Tomasz Kacprzak	ETH Zurich	DES+DESC Generative AI for cosmology
Yihui Ren	Brookhaven National Laboratory	Feasibility study on AI Foundation Models for Particle Physics and Accelerators
Da Yang	University of Chicago	Constructing a hierarchy of ML models to understand the physics of the Madden-Julian Oscillation
Tommi Jaakkola	MIT	Structural Token Language Modeling of Proteins for Controllable Design
Rafael Gomez Bombarelli	MIT	Generative AI for Rigorous Thermodynamics of Material
Shinjae Yoo	Brookhaven National Laboratory	Foundation Neuroscience AI Model (NeuroX)
Pei Zhang	Oak Ridge National Laboratory	Multiscale foundation models for computational fluid dynamics
Tim Smith	NOAA Physical Sciences Laboratory	Ensemble Weather Generators for Data Assimilation and Forecasting at NOAA
Pablo Unzueta	Stanford University	Accelerating quantum chemistry and ab initio molecular dynamics with generative AI and deep learning
Aditi Krishnapriyan	University of California Berkeley	Generative Modeling for Molecular Dynamics using Graph Latent Differential Equations
Christopher Sutton	University of South Carolina	Advancing Heterogeneous Catalyst Design through Generative AI
Sokrates Pantelides	Vanderbit University	Unraveling vibrational properties of nanoconfined water through density functional theory and deep learning approaches
Mingda Li	MIT	Advanced Diffusion Model Strategy for Identifying Quantum Spin Liquid Candidates
Bin Yu	University of California Berkeley	Low Rank Fine-Tuning for One-Step to Many-Step Skill Transfer: Applying LoRA+ to Improve Medium-Range and Long-Range Weather and Climate Skill
Radi Radev	CERN	Scalable Sparse Diffusion Models for Neutrino Physics
Kristopher Bouchard	Lawrence Berkeley National Laboratory	Foundation Generative Pretrained Model for Proteins
Rajni Chahal	Oak Ridge National Laboratory	Uncertainty-guided Machine Learning Interatomic Potentials for Reactive Polymers
Rose Yu	University of California San Diego	Efficient diffusion-based modeling for weather and climate prediction
Shengchao Li	University of California Berkeley	Foundation Model on Materials
Weiqiang Zhu	University of California Berkeley	Foundation Model for Seismic Data

About NERSC and Berkeley Lab
The National Energy Research Scientific Computing Center (NERSC) is a U.S. Department of Energy Office of Science User Facility that serves as the primary high performance computing center for scientific research sponsored by the Office of Science. Located at Lawrence Berkeley National Laboratory, NERSC serves almost 10,000 scientists at national laboratories and universities researching a wide range of problems in climate, fusion energy, materials science, physics, chemistry, computational biology, and other disciplines. Berkeley Lab is a DOE national laboratory located in Berkeley, California. It conducts unclassified scientific research and is managed by the University of California for the U.S. Department of Energy. »Learn more about computing sciences at Berkeley Lab.