Request for Proposals (RFP)
RFP for GenAI4Science Seed Projects
(Deadline: Monday, November 10, 2025)
Link to GenAI Mixer slides with useful links and resources.
Introduction & Purpose
The CSE Data Science Initiative (DSI) invites proposals for seed projects in the emerging area of Generative AI for Science and Engineering (GenAI4Science).
Generative AI is poised to become a central driver of progress across many fields of science and engineering. Just as advances in vision and language modeling reshaped AI in the last decade, the next breakthroughs in GenAI will be catalyzed by challenges from scientific and engineering domains where deep knowledge has been accumulated over decades and even centuries. AI is also one of the areas that will be invested in most heavily by the current administration, creating unprecedented opportunities for leadership and funding. As highlighted in the America’s AI Action Plan, “Just as LLMs and generative AI systems represented a paradigm shift in the science of AI, future breakthroughs may similarly transform what is possible with AI.”
For CSE — a college that unites expertise across science and engineering with leadership in foundational AI — collaboration at this interface is not only timely but essential. Our success as a college depends on harnessing GenAI4Science to both accelerate discovery in the natural and applied sciences and advance the foundations of AI itself.
Building on the momentum of the GenAI4Science Day (May 20, 2025) and the GenAI4Science Workshop (August 13–14, 2025) held at UMN, these seed projects aim to advance both scientific discovery and foundational GenAI methods, with a strong emphasis on the two central pillars of the White House AI Action Plan: AI for Science and the Science of AI. You can find additional useful science policy documents addressing the future of AI and GenAI on our Science Policy webpage.
Collaboration Opportunities
To facilitate the formation of research teams, CSE DSI will host two GenAI4Science Mixers this fall:
- Friday, October 3, 2025
- Friday, October 17, 2025
Location: Keller 3-180. Lunch will be provided at noon; the program runs 12:30–2:30 pm.
Please use this form to sign up and volunteer to give a lightning talk. Affiliates may also prepare or update quad charts to help identify potential collaborators (contact Bobbie at [email protected] for assistance).
Please sign up by Friday, September 26, to aid us in our planning.
Proposal Guidelines
Proposed projects should form collaborative research teams involving faculty from at least two departments, with participation from both:
- A domain science or engineering discipline (e.g., physical sciences, life sciences. health, environmental sciences, and agriculture, ), and
- A data science/AI discipline (e.g., computer science, electrical engineering, mathematics, industrial engineering, and statistics).
Projects should clearly demonstrate how expertise from different disciplines connects through a common scientific challenge or through collaboration toward a shared GenAI discovery, invention or solution.
Selection Priorities
Proposals must articulate how the work has the potential to grow into a competitive proposal to external sponsors (government, industry, foundations). Priority will be given to ideas that:
- Advance novel GenAI methods by integrating scientific knowledge into model design, training, or evaluation.
- Address domain challenges of high societal impact (climate, health, food and agriculture, manufacturing, materials) with principled choice of GenAI methods (e.g., autoregressive, diffusion, GANs, normalizing flows).
- Develop benchmarks, datasets, or testbeds that can serve as community resources.
- Position UMN to lead in national initiatives and federal investments in GenAI4Science, aligning with the White House AI Action Plan and the federal emphasis on AI investment.
Funding & Support
We expect to fund up to six collaborative teams, with each award providing one graduate research assistantship for one year (20 hrs/week, starting Spring 2026). Each assistantship must be jointly advised by the faculty team members and will be open to graduate students enrolled in CSE. Assistantships may be split between two CSE students.
The goal is to enable projects to develop preliminary results that strengthen a larger external proposal.
Proposal Format (3 pages max, including graphs/figures and references)
Submit your proposal as a PDF file to Bobbie at [email protected] by Monday, November 10, 2025.
Each proposal should explicitly include:
- Innovativeness of the proposed GenAI4Sc. work for AI for Science and the Science of AI.
- Scope of proposed research with specific hypotheses, likely outcomes, tasks and milestones along with datasets (either available or to be produced or both).
- Team qualification as well as roles and responsibilities
- Graduate student role (how the assistantship will enable progress).
- Growth potential — including relevant sponsors/programs (e.g., NSF, USDOD, USDOE, NASA, NIH, USDA, industry, foundations) and alignment with America’s AI Action Plan.
Review Process & Evaluation Criteria
Proposals will be reviewed by a panel of CSE faculty with expertise spanning AI, data science, and domain sciences/engineering. Reviews will consider the following criteria:
- Alignment with GenAI4Science vision : Clear connection to one or both pillars of the White House AI Action Plan (AI for Science and Science of AI).
- Innovation: Novelty of GenAI methods and/or integration of scientific knowledge into AI.
- Collaboration & interdisciplinarity: Strength of cross-departmental team spanning domain science/engineering and data science/AI.
- Impact: Potential to address challenges of high societal importance and position UMN as a leader.
- Growth potential: Likelihood of leading to competitive external proposals (e.g., NSF, DOE, NIH, USDA, foundations, industry).
Funding decisions will be announced in December 2025, with projects beginning in Spring 2026.