Project Grant 2611549
- The National Science Foundation Division of Computing and Communication Foundations awarded the University of Wisconsin - Madison $451,577 on August 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop approximate Bayesian inference methods for diffusion posteriors in generative artificial intelligence. The project addresses a gap in current generative models used for scientific applications: while these models can produce realistic outputs, they...
- This $600,000 project grant, awarded by the National Science Foundation (NSF) under the Computer and Information Science and Engineering (CFDA 47.070) program, will support research to develop a new Bayesian diffusion model framework for advanced visual perception and cognition systems. The University of California, San Diego (UCSD) will serve as the primary awardee, with the goal of revisiting the analysis-by-synthesis methodology by integrating generative priors into the learning and inference...
- This five-year Project Grant award of $446,455 from the National Science Foundation's Division of Information and Intelligent Systems supports research in variational inference for resource-efficient learning under the agency's Computer and Information Science and Engineering program (CFDA 47.070). The University of California Irvine will conduct investigator-initiated research from September 2021 through August 2026 to advance the development of variational inference techniques that enable more...
- The National Science Foundation Division of Mathematical Sciences awarded the University of California Irvine $209,660 on August 1, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) to develop mathematical and artificial intelligence methodologies for wave-based imaging technologies. The recipient will establish rigorous mathematical foundations for generative AI techniques applied to imaging and uncertainty quantification in infinite-dimensional settings. The work focuses...
- The National Science Foundation Division of Electrical, Communications and Cyber Systems awarded the University of California, Berkeley $150,000 on September 1, 2026, under the Engineering program (CFDA 47.041) to develop mathematical foundations for reinforcement learning and generative artificial intelligence systems. The project will establish stochastic analysis, stochastic control, and optimization theory for three interconnected research directions: continuous-time reinforcement learning...
- The National Science Foundation Division of Computing and Communication Foundations awarded $384,398 to the Regents of the University of Michigan on August 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) for theoretical and algorithmic research on diffusion models in generative artificial intelligence. The project, titled "Towards Theoretical and Algorithmic Foundations of Diffusion Models," aims to establish mathematical and computational...
- The National Science Foundation Division of Computing and Communication Foundations awarded the Regents of the University of California at Riverside $300,000 on September 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop uncertainty quantification methods for large language models used in high-stakes reasoning and decision-making. The project, performed in Riverside, California, through August 31, 2029, addresses the core problem of language...
- The National Science Foundation Division of Mathematical Sciences awarded the University of California Irvine $299,999 on August 15, 2026, under the Mathematical and Physical Sciences program (CFDA 47.049) to conduct theoretical research on discrete approximation, analyzing functions of many yes/no variables and their continuous analogues in discrete and Gaussian spaces. The project investigates how much information is required to learn low-complexity functions, how well complicated functions...
- The National Science Foundation Division of Information and Intelligent Systems awarded the Regents of the University of California at Riverside $250,000 on August 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to support collaborative research on SHF: SENSABRIDGE, a learnable sensor-AI bridge for ubiquitous edge intelligent processing. Work will be performed in Riverside, California, with a period of performance extending through July 31, 2029. No...
- The National Science Foundation Division of Information and Intelligent Systems awarded the University of California, San Diego $215,035 on May 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) for collaborative research on compiler infrastructure redesign to support scientific computing. The project develops methods to synthesize domain-specific and program-specific static analyses that enable scientists to express and statically enforce complex...
The National Science Foundation Division of Computing and Communication Foundations awarded the University of California Irvine $548,422 on August 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop approximate Bayesian inference methods for diffusion posteriors. The research creates posterior sampling methods that integrate pretrained diffusion models with observed data, making generative artificial intelligence systems more reliable for scientific applications. The work addresses a critical gap: current generative models often fail to quantify uncertainty or faithfully preserve scientific constraints when used in medical imaging, weather forecasting, satellite image analysis, molecular discovery, and similar high-stakes domains. The project develops unbiased gradient estimators, including reinforcement and reparameterization approaches, to improve how these models handle observations and represent uncertainty in settings with limited data. Deliverables include open-source software, trained models, and benchmarks for public release, alongside educational components spanning summer schools, workshops, student research opportunities, and citizen science activities using community-contributed data. Performance occurs at the University of California Irvine in Irvine, California. The period of performance runs from August 1, 2026, through July 31, 2030. This is a collaborative NSF-BSF (National Science Foundation–Binational Science Foundation) project.
Mod # | Description | ReasonForModification | Federal Obligation | Date |
|---|---|---|---|---|
| Not listed | $548.4k | 7/28/26 |