Does the Explanation Drift Before the Guarantee? Label-Free Early Warning on Evolving Graphs

Key points

• Deployed models can fail silently as the world they describe evolves.1
• A vanishing edge loses mask mass regardless of whether the model relied on it, so we condition the importance on edge presence.1
• Evolving graphs also break edge identity: a relabelling that changes nothing shifts the identity-based statistic by 0.69 on synthetic graphs and 0.83 on real ego-networks, whereas aggregating over structural roles brings both to 0.0 and stays well-defined when the support changes.1
• The main limitation is that the mass channel is blind to covariate drift, which breaks the guarantee in every run; the distributional channel detects it and also rescues the rewiring regime, so we report the two channels separately rather than as a union.1

Sources (1)

• [1]Does the Explanation Drift Before the Guarantee? Label-Free Early Warning on Evolving Graphs
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 12, 10:22 AM

  Attribution drift has served as an early warning in tabular data; we ask whether, in graph streams, a model's own explanations can forewarn the loss of nominal conformal coverage.

  Deployed models can fail silently as the world they describe evolves.

Extractive summary: sentences quoted from the sources.