2021/05/18 by Suzanne Kawamleh · 14 citations
Arts and Humanities · Psychology · Social Sciences · #Algorithm #Artificial intelligence #Artificial neural network #Climate Change Communication and Perception #Climate change #Climate model #Climate science #Computer science #Ecology #Epistemology #Generalizability theory #Machine learning #Management science #Mental Health Research Topics #Parameterized complexity #Philosophy and History of Science #Political science #Politics #Power (physics) #Process (computing) #Psychology #Reliability (semiconductor) #Representation (politics)
paper · doi:10.1086/714877
published in Philosophy of Science 88(5), 1008-1020 (Cambridge University Press)
openalex publication_date 2021/05/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/07
Scientists and decision makers rely on climate models for predictions concerning future climate change. Traditionally, physical processes that are key to predicting extreme events are either directly represented (resolved) or indirectly represented (parameterized). Scientists are now replacing physically based parameterizations with neural networks that do not represent physical processes directly or indirectly. I analyze the epistemic implications of this method and argue that it undermines the reliability of model predictions. I attribute the widespread failure in neural network generalizability to the lack of process representation. The representation of climate processes adds significant and irreducible value to the reliability of climate model predictions.