{"id":"arxiv/2603.15431","url":"https://arxiv.org/abs/2603.15431","tags":["cs.LG","cs.AI","math.AP","math.NA"],"score":0,"title":"Physics-informed fine-tuning of foundation models for partial differential equations","domain":"cs.LG","source":"arxiv","authors":["Vlad Medvedev","Leon Armbruster","Christopher Straub","Georg Kruse","Andreas Rosskopf"],"reviews":0,"subject":"cs.LG","summary":"Foundation models for partial differential equations (PDEs) have emerged as powerful surrogates pre-trained on diverse physical systems, but adapting them to new downstream tasks remains challenging due to limited task-specific data and distribution shifts. While fine-tuning has proven transformative in natural language processing, best practices for adapting PDE foundation models remain underexplored. Although physics-informed training has successfully trained accurate solvers across a wide range of PDE problems, its potential for fine-tuning data-based foundation models has not been…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-03-16","reactions":{}}