CSE DSI GenAI Day and Open House

Update

Recordings of several of the morning talks are now posted on the CSE DSI YouTube Channel on the GenAI Day 2025 Playlist.

#####

The Generative AI (GenAI) revolution—exemplified by the remarkable success of foundation models and applications like ChatGPT that leverage these models—is beginning to see widespread adoption across science and engineering. Specifically, foundation models (i.e., pre-trained models capable of being fine-tuned on a wide variety of tasks) are now being used across nearly all scientific disciplines. In some cases, these models already outperform traditional methods employed by their respective scientific communities, highlighting their transformative potential in driving scientific discovery.

The College of Science and Engineering (CSE) Data Science Initiative (DSI) is holding a Generative AI (GenAI) Day and Open House on May 20 from 10:00 to 3:00 in Keller Hall. This event will serve as a starting point for building communities around GenAI and exploring its transformative potential across disciplines. It will also highlight some of the exciting research being pursued by our affiliates and showcase our new interaction space in Keller Hall.

This event will be the first step in a series of activities we are planning to enable UMN to take leadership in this fast-evolving area of great importance. Future activities will support: 

1. Evaluating the potential of generative AI to drive scientific discovery while assessing its limitations (e.g., hallucinations) and exploring approaches to enhance its reliability and effectiveness (e.g., by incorporating scientific theories);
2. Identifying gaps in the current capabilities of generative AI for addressing grand scientific challenges and outlining the AI advancements needed to foster synergistic research across scientific disciplines;
3. Identifying actionable strategies for effectively integrating these advancements into the practice of scientific discovery.

Agenda

We will add more detailed information as it becomes available. 

10:00am - noon in Keller 3-210

10:00: Welcome and Introductions

10:15 - 10:55: Introduction to Generative AI in Science and Engineering with discussion

Aryan Deshwal (CS&E) and Anuj Karpatne (Computer Science, Virginia Tech)

11:00 - noon: Panel and discussion on GenAI challenges and opportunities in science & engineering. 

Panel moderator: Anuj Karpatne

Serguei Pakhomov (Pharmaceutical Care & Health Systems)
Ellad Tadmor (AEM)
Yao-Yi Chiang (CS&E)
Lucy Fortson (School of Physics & Astronomy)
Stevie Chancellor (CS&E)

Noon - 2:00pm in Keller 3-210, 3-111, 3-115, 2-260

Noon - 12:15  Pick up lunch, go to breakout session

12:15 - 1:15   Breakouts 

Each breakout will have two conveners to guide the discussion. The themes below were informed by the recent report from the 2nd NSF-Sponsored Workshop on the AI-Enabled Scientific Revolution, which focused on generative AI. 

Keller 3-210: Uncertainty Quantification (UQ) and new directions enabled/motivated by GenAI to address it (including statistics, faster Monte Carlo simulations).
Conveners: Jie Ding (School of Statistics), Nadja Strobbe (School of Physics & Astronomy), Mingyi Hong (ECE)

Keller 3-111: Generalization on out of distribution data and handling rare classes (including approaches such as incorporating domain knowledge, human-AI collaboration).
Conveners: Jarvis Haupt (ECE), Yoga Varatharajah (CS&E)

Keller 3-115: Limitations in experiment design (including approaches such as Bayesian experimental design, data-driven hypothesis generation, automation to speed up decision making).
Conveners: Aryan Deshwal (CS&E), Sapna Sarupria (Chemistry)

Keller 2-260: Inability to engage in symbolic reasoning, decision making, and explanations (including approaches such as augmenting GenAI with complementary AI paradigms).
Conveners: Yao-Yi Chiang (CS&E), Martin Michalowski (School of Nursing)

1:30 - 2:00   Summaries in Keller 3-210

2:00 - 3:00pm in Keller 4-250

Open House

Receptions and posters in CSE DSI interaction space in and around Keller 4-250.

Useful Resources

Draft report from 2nd NSF-Sponsored Workshop on the AI-Enabled Scientific Revolution held in Minneapolis in August 2024.

Generative Artificial Intelligence in Health and Medicine: Opportunities and Responsibilities for Transformative Innovation  From the National Academy of Medicine, 2025.