Id: arxiv/2611.13566 Url: https://arxiv.org/abs/2611.13566 Tags: cs.LG, cs.AI, eess.SY, math.OC Score: 0 Title: Metamorphic testing of learned AC optimal power flow surrogates reveals representation dependence Domain: cs.LG Source: arxiv Authors: Stephen Peng, Vineet Jagadeesan Nair Reviews: 0 Subject: cs.LG Category: research Comments: 0 Disputed: false Division: cs Mentions: 0 Discussed: 0 Published: 2026-11-10 Page: https://delphi.science/r/arxiv/2611.13566 Evaluating a learned AC optimal power flow (AC-OPF) surrogate on one grid serialization cannot show whether its accuracy depends on that encoding. We test this dependence with thirteen metamorphic relations, each specifying when a transformation is valid and how to compare its outputs. Our audit covers 17 model and pipeline configurations: two released trained checkpoints and 15 random-weight architecture or mechanism controls. Every configuration violates at least six relations, even at reporting tolerances ten times tighter or looser. For GridSFM-Open v1.1, source inspection and…