Continental Breakfast (provided)

Presenter: Shaowen Wang, University of Illinois Urbana-Champaign

• From Manual GIS Workflows to Reproducible Dam-Failure Risk Analytics: a PostGIS REST API and Dashboard Reference Implementation. Jungha Woo (Purdue University); Xiao Liu (Purdue University); Erick Li (University of Illinois Urbana-Champaign); Michael Englert (Utah State University); Furqan Baig (University of Illinois Urbana-Champaign); Carol Song (Purdue University); Anand Padmanabhan (University of Illinois Urbana-Champaign); Courtney Flint (Utah State University)
• A Convergence Science Approach to Fertilizer Market Shocks: Integrating Geospatial, Economic, and Environmental Dynamics. Iman Haqiqi (Purdue University)

• MURD-ViT: Urban Retrofitting Detection with Vision Transformer. Raj Bhattarai (Virginia Tech University) and Fangzheng Lyu (Virginia Tech University)
• GeoLocate: Spatial Modeling of Market Entry Viability. Jaiany Rocha Trindade (Federal University of Rio Grande do Sul, Brazil & Center for Geographic Analysis, Harvard University), Vinicius Andrade Brei (Federal University of Rio Grande do Sul & MIT Media Lab), and Devika Jain (Center for Geographic Analysis, Harvard University)
• From Mobility Intensity to Market Infrastructure: A Spatial AI Framework for Recovering Hidden Spatial Regimes and Multiscale Variation. Rafael Albuquerque (Federal University of Rio Grande do Sul, Brazil), Jéssica Miranda (Federal University of Rio Grande do Sul, Brazil), Vinicius Andrade Brei (Federal University of Rio Grande do Sul & MIT Media Lab), and Siqin (Sisi) Wang (Spatial Sciences Institute, University of Southern California)

• Evaluating Storage and I/O Efficiency of Geospatial Data Formats. Fangzheng Lyu (Virginia Tech University)
• QGLLM: Quant-Generative Large Language Model Framework for Sea Ice Prediction over the Arctic. Shanmukhi Gundu (iHARP, University of Maryland, Baltimore County); Sudip Chakraborty (iHARP, University of Maryland, Baltimore County); Vandana Janeja (iHARP, University of Maryland, Baltimore County)
• Evaluating Alternative Designs of the 30×30 Biodiversity Target in a Spatially Explicit Global Economic Model. Iman Haqiqi (Purdue University)

        Presenter: Luc Anselin, Stein-Freiler Distinguished Service Professor of Spatial Data Science, University of Chicago

Continental Breakfast (provided)

Presenter: Chaitan Baru, Emerging Technologies in the Technology, Innovation, and Partnerships (TIP) Directorate, National Science Foundation (NSF)

• • Yuqin Jiang, University of Hawaii at Manoa
  • Debayan Mandal, Arizona State University
  • Fangzheng Lyu, Virginia Tech University
  • Eric Shook, University of Minnesota

Guiding question: What should AI-era cyberinfrastructure education and workforce development actually look like?

Talk (20 mins):

• AI Is Changing Geographic Problem Solving; Instruction Must Change Too. Peter Kedron (University of California Santa Barbara); Sarigai Sarigai (University of California Santa Barbara); Heather Baier (Arizona State University); Andrew Trgovac (Arizona State University)

Position statements (5 mins each):

• Leverage the National Student Data Corps for Multi-domain HDR AI Workforce Development. Florence Hudson (Columbia University); Emily Rothenberg (Columbia University)
• Toward an AI-HPC Skills Inventory: Effectively Assessing Advanced Cyberinfrastructure Upskilling Interventions at Research Institutions. Roy Prouty (University of Maryland, Baltimore County)
• Education and Workforce Development for AI-enabled Sciences. Amin Sahba (The University of Texas at San Antonio

Breakout groups (40 mins):

• Group A — Curriculum & instruction for GeoAI (Lead: Peter Kedron)
• Group B — Community & workforce pipelines (Lead: Florence Hudson)
• Group C — Measuring upskilling impact (Lead: Roy Prouty)
• Group D — Education & workforce for AI-enabled science (Lead: Amin Sahba)

Report backs  (15 mins)

Continental Breakfast (provided)

Guiding Question: What does trustworthy AI-enabled science actually require in practice?

Position Statements (5 minutes each)

• • Rethinking Location Privacy: Cross-Modal Inference in Foundation Models. Yue Lin (University of Illinois Urbana-Champaign)
  • Uncertainty Quantification in Heterogeneous Data Environments. Jun Wu (Michigan State University)
  • Compliance by Design: Embedding Responsibility in AI Cyberinfrastructure. Peter Darch (University of Illinois at Urbana-Champaign)
  • Reproducibility in Scientific Workflows with Artificial Intelligence: Challenges and a Path Forward. Smita Darmora (University of Illinois Chicago & College of DuPage)
  • Ethical and Responsible AI. Ramin Sahba (The University of Texas at San Antonio)
    

Breakout Group Discussions (45 minutes)

• • Group A — Privacy & inference risk (Lead: Yue Lin)
  • Group B — Uncertainty & reliability (Lead: Jun Wu)
  • Group C — Accountability & compliance (Leads: Peter Darch, Ramin Sahba)
  • Group D — Reproducibility of AI Workflows (Lead: Smita Darmora )

Report Backs (20 minutes)

Guiding Question: How do we build AI that generalizes and holds up under real-world scenarios?

Position Statements (5 minutes each)

• Moving Beyond Correlation: Causal AI for Generalizable Earth Sciences. Emam Hossain (University of Maryland, Baltimore County) and Md Osman Gani (University of Maryland, Baltimore County)
• Time Critical Mission Support: Opportunities for AI, HPC, Geospatial Digital Twins. Budhendra Bhaduri (Oak Ridge National Laboratory)
• From Hazard to Opportunity: A Convergence Framework for Wildfire Disaster Prediction and Resilience. Janice Coen (NSF NCAR)
• Identifying Cryospheric Connectivity: a Graph-Based Framework Linking Sea Ice Retreat to Ice Shelf Melting. Molly Balkan (iHARP, University of Maryland, Baltimore County); Shanmukhi Gundu (iHARP, University of Maryland, Baltimore County); Sudip Chakraborty (iHARP, University of Maryland, Baltimore County); Vandana Janeja (iHARP, University of Maryland, Baltimore County)
• AI for Science and Science for AI: What Happens When the Data Aren't There? Elif Ertekin (University of Illinois Urbana-Champaign)

Breakout Group Discussions (45 minutes)

• Group A — Causal & generalizable Earth AI (Leads Emam Hossain and Elif Ertekin)
• Group B — Time-critical & disaster support (Leads: Budhendra Bhaduri and Janice Coen )
• Group C — Earth system dynamics (Lead: Molly Balkan )

Report Backs (20 minutes)

Discussion topics will be solicited from attendees throughout the day via a shared topic board, where anyone can propose questions sparked by discussions throughout the week. The afternoon opens with a rapid pitch-and-cluster, moves into self-organized concurrent topic groups (unconference style), and closes with report outs.

Throughout the day.  Open topic board collects proposed discussion questions from participants. Participants can also vote on topics that would be selected. 

Pitch. Proposers give 2-3 min  topic pitches; related topics grouped into tables (15 mins)

Unconference. Concurrent self-organized topic groups; participants move freely between tables (60 mins)

Table leads report back. (5 minutes each)