Tools for the Tool-Builders: LLM-Assisted Software, Algorithms, and AI/ML in the Geodynamic Modelling Community
Tools for the Tool-Builders: LLM-Assisted Software, Algorithms, and AI/ML in the Geodynamic Modelling Community
Session ID#: 282209
Session Description:
This session is for researchers, software engineers, algorithm developers and data analysts whose working lives have been upended by AI programming tools and machine-learning algorithms, and for those watching from the sidelines and wondering what to make of it. Whether you've leaned in, held back, or are still making up your mind, we want your experiences. The geodynamic and seismological modelling community depends on collaboratively developed open-source software, and the way that software is written, maintained, taught, and trusted is quietly being rewritten around us.
We invite contributions on productivity gains, false starts and cautionary tales, and novel approaches — across software engineering, algorithm design and implementation, and AI/ML as scientific tools in their own right.
The worries: reproducibility, validation, provenance, attribution, code ownership, environmental cost; and a harder question: does AI access help smaller groups punch above their weight, or entrench the advantages of those already at the frontier?
Co-Sponsor(s):
- DI - Study of the Earth's Deep Interior
- IN - Informatics
- S - Seismology
Index Terms:
1906 Computational models, algorithms [INFORMATICS]
1942 Machine learning [INFORMATICS]
1956 Numerical algorithms [INFORMATICS]
1976 Software tools and services [INFORMATICS]
Primary Convener: Louis Noel Moresi, Australian National University, Research School of Earth Sciences, Canberra, ACT, Australia
Conveners: Alice-Agnes Gabriel, Ludwig-Maximilians-Universität München, Department of Earth and Environmental Sciences, Munich, Germany, Lorraine Hwang, University of California, Davis, Davis, CA, United States and Jed Brown, Argonne National Laboratory, Argonne, United States
See more of: Tectonophysics