aiqbench analysis
This research addresses the critical interpretability gap in deep learning models used for predicting extreme weather events like fires and cyclones. By integrating geographic modulation with rule-based Sparse Autoencoders, the study successfully translates complex multi-modal climate data into human-understandable rules that align with scientific literature. This advancement is significant for operational settings where trust and explainability are prerequisites for deploying AI-driven climate models.
SAE-Xplainers: Rule-Based Feature Interpretation for Extreme Earth Events
arXiv:2608.20117v1 Announce Type: new Abstract: The emergence of large-scale Weather and Climate (W&C) datasets offers new opportunities for modeling extreme Earth events (ExEE) and their impacts using deep learning. However, their adoption in operational settings remains limited by the lack of models' interpretability. While for conventional text and image modalities, tools such as Sparse Autoencoders (SAEs) have proven effective for extracting human-understandable concepts, their use for the analysis of ExEE remains challenging due to the nature of W&C data. To address this, we introduce (i) a geographic location-based modulation of the inputs of SAE to capture the local semantic meaning of environmental patterns, and (ii) an ensemble of rule-based SAE-Xplainers to interpret the resulting high-dimensional features derived from complex, multi-modal environmental predictors. We evaluate our method on three ExEE types: the prediction of fires, and the detection of tropical cyclones and atmospheric rivers. We show that SAE input modulation improves both reconstruction performance and feature utilization, and that our SAE-Xplainers enable faithful interpretation of complex climatic patterns by unfolding them into human-understandable rules that are consistent with the scientific literature, while also supporting the identification of feature absorption.
Abstract:The emergence of large-scale Weather and Climate (W&C) datasets offers new opportunities for modeling extreme Earth events (ExEE) and their impacts using deep learning. However, their adoption in operational settings remains limited by the lack of models' interpretability. While for conventional text and image modalities, tools such as Sparse Autoencoders (SAEs) have proven effective for extracting human-understandable concepts, their use for the analysis of ExEE remains challenging due to the nature of W&C data. To address this, we introduce (i) a geographic location-based modulation of the inputs of SAE to capture the local semantic meaning of environmental patterns, and (ii) an ensemble of rule-based SAE-Xplainers to interpret the resulting high-dimensional features derived from complex, multi-modal environmental predictors. We evaluate our method on three ExEE types: the prediction of fires, and the detection of tropical cyclones and atmospheric rivers. We show that SAE input modulation improves both reconstruction performance and feature utilization, and that our SAE-Xplainers enable faithful interpretation of complex climatic patterns by unfolding them into human-understandable rules that are consistent with the scientific literature, while also supporting the identification of feature absorption.
22 pages, 16 Figures, Under Review. A non-archival 2-page version was accepted as an oral presentation at Climate Informatics 2026 (Extended Abstract ID 66, this https URL), and a non-archival 4-page version was accepted as an oral presentation at the AICC Workshop at ECCV 2026
Machine Learning (cs.LG)
arXiv:2608.20117 [cs.LG]
(or arXiv:2608.20117v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2608.20117 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 246 words
