Aggregation strategies to improve XAI for geoscience models that use correlated, high-dimensional rasters

dc.contributor.authorKrell, Evan
dc.contributor.authorKamangir, Hamid
dc.contributor.authorCollins, Waylon
dc.contributor.authorKing, Scott A.
dc.contributor.authorTissot, Philippe
dc.date.accessioned2024-02-14T19:01:32Z
dc.date.available2024-02-14T19:01:32Z
dc.date.issued2023-10-30
dc.description.abstractComplex machine learning architectures and high-dimensional gridded input data are increasingly used to develop highperformance geoscience models, but model complexity obfuscates their decision-making strategies. Understanding the learned patterns is useful for model improvement or scientific investigation, motivating research in eXplainable artificial intelligence (XAI) methods. XAI methods often struggle to produce meaningful explanations of correlated features. Gridded geospatial data tends to have extensive autocorrelation so it is difficult to obtain meaningful explanations of geoscience models. A recommendation is to group correlated features and explain those groups. This is becoming common when using XAI to explain tabular data. Here, we demonstrate that XAI algorithms are highly sensitive to the choice of how we group raster elements. We demonstrate that reliance on a single partition scheme yields misleading explanations. We propose comparing explanations from multiple grouping schemes to extract more accurate insights from XAI. We argue that each grouping scheme probes the model in a different way so that each asks a different question of the model. By analyzing where the explanations agree and disagree, we can learn information about the scale of the learned features. FogNet, a complex three-dimensional convolutional neural network for coastal fog prediction, is used as a case study for investigating the influence of feature grouping schemes on XAI. Our results demonstrate that careful consideration of how each grouping scheme probes the model is key to extracting insights and avoiding misleading interpretations.
dc.description.sponsorshipThis research was supported by grants from the National Science Foundation (2019758) and the National Science Foundation (1828380).
dc.identifier.citationKrell E, Kamangir H, Collins W, King SA and Tissot P (2023). Aggregation strategies to improve XAI for geoscience models that use correlated, high-dimensional rasters. Environmental Data Science, 2: e45. doi:10.1017/eds.2023.39
dc.identifier.doihttps://doi.org/10.1017/eds.2023.39
dc.identifier.urihttps://hdl.handle.net/1969.6/97753
dc.language.isoen_US
dc.rightsCC BY 4.0 DEED Attribution 4.0 International
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/
dc.subjectcoastal fog forecasting
dc.subjectdeep learning
dc.subjectexplainable artificial intelligence
dc.subjectgeospatial raster data
dc.titleAggregation strategies to improve XAI for geoscience models that use correlated, high-dimensional rasters
dc.typeArticle

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
Aggregation strategies to improve XAI for geoscience models that use correlated, high-dimensional rasters.pdf
Size:
12.1 MB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.72 KB
Format:
Item-specific license agreed upon to submission
Description: