3A Data Quality and Provenance for Artificial Intelligence (AI) and Machine Learning (ML) Applications

Monday, 29 January 2024: 1:45 PM-3:00 PM
337 (The Baltimore Convention Center)
Host: 40th Conference on Environmental Information Processing Technologies
Cochairs:
Shawn W. Miller, NASA, 474, Aurora, CO and Eric P. Grimit, NCAR, Research Applications Laboratory, Boulder, CO

  1. As Artificial Intelligence (AI) and Machine Learning (ML) usage in weather, water, and climate applications increases, the quality and provenance of data used for model training and application are also evolving to become increasingly important. We explored tough questions with a panel session at the 2023 AMS Washington Forum, such as: What ground truth is used in model training? Are the limitations of ground truth data considered? The Annual Meeting provides an opportunity to continue and add further definition and guidance to the conversation. Specific items to be considered include, but are not limited to: 1) What kinds of filtering, processing, or transformations are applied to the data between original observations or model runs and ingest/usage by the AI/ML user, and where these changes are best applied; 2) Measures of quality and context surrounding a given data set, including whether the associated metadata are sufficiently complete to understand these aspects; 3) Measures of the appropriateness for a given type of data to solve a given type of problem. This session will explore how we ensure traceability and selection of the best data possible for a given AI/ML application to maximize accuracy and effectiveness to the respective end users.

Papers:
1:45 PM
3A.1
Introduction to Development of Meteorological Big Data Preprocessing Technology for Artificial Intelligence Training Data Production
Junsang Park, National Institute of Meteorological Sciences, Korea Meteorological Administration, Deogwipo, South korea; KMA, Seogwipo, South korea; and K. Park, H. Lee, S. Hong, J. Kim, and G. LEE

2:00 PM
3A.2
Efficient and Rigorous Data Quality Checking for Training a Machine Learning Postprocessing Algorithm on High-Resolution NWP Data and Surface Observations
Ashley Elizabeth Payne, Tomorrow.io, Golden, CO; and K. Keshavamurthy, L. Conibear, A. Reed Harris, T. McCandless, M. E. Green, MA, and S. Flampouris

2:15 PM
3A.3
Leveraging Machine Learning and Deep Learning for Enhanced Data Quality Control at the Atmospheric Radiation Measurement (ARM) User Facility
Mia Li,

Recorded Presentation
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
See more of: 40th Conference on Environmental Information Processing Technologies
<< Previous Session | Next Session >>