Seeing the smoke before the fire

Note: To view the web version of this presentation with working animations visit (https://janithwanni.quarto.pub/seeing-the-smoke-before-the-fire/)

Abstract: The prevailing reliance on high-performance black box models calls for a shift towards Explainable AI (XAI) methodologies. Anchors, a subset of Explainable AI (XAI) methods, work by generating a bounding box around a given local instance that contain observations with similar predictions. Exploring the anchors obtained from explaining spatio-temporal black box models provides us with the ability to understand and explore the underlying relationships through space and time.

We introduce high-dimensional data visualisation methods designed to understand explanations derived from black box models, thereby enabling the extraction of latent features and decision pathways across temporal and spatial dimensions. This work is motivated by the need to advance bushfire management strategies with the aim of preventing, mitigating and provide tools for managing future catastrophic bushfires similar to those in 2019-2020. By using early detection data from hotspots combined with a holistic understanding of the scenarios leading to higher fire ignition risk this work aims at contributing to the bush fire risk management.