Back to AI News
arxiv

Pricing Symbolic Rules by What GNNs Leave Unlearned

This paper proposes an information-theoretic framework for evaluating symbolic rules added to graph neural networks, measuring the exact log-loss reduction a rule can provide given what the model family computes. The authors show that rule design can be treated as rule discovery, scoring candidates on the network's out-of-fold residual against calibrated permutation nulls. On ZINC, the method recovers a hidden ring term and beats expert cycle counts, and on BBBP it significantly outperforms both the plain network and the expert rule.

AI Brief

Provides a principled, exact framework for symbolic rule discovery in GNNs with strong empirical results, offering practical value for model interpretability and improvement.

arXiv:2611.12446v1 Announce Type: cross Abstract: Symbolic knowledge (motif counts, ring counts, label constraints) is routinely added to graph neural networks (GNNs) and justified by their limited expressivity. We replace this yardstick with an exact one. The largest log-loss reduction a rule $S$ can give a model family is $I(Y;S\mid\sigma)$ given what the family computes; every realised gain splits exactly into this term plus repaid neural debt; and for a trained network with representation $T$ the value is $I(Y;S\mid T)$, the label information training failed to extract. Rule design thus becomes rule discovery: we score candidates on the network's out-of-fold residual against calibrated permutation nulls with family-wise error control. On ZINC, discovery recovers the hidden ring term $\max(0,\ell_{\max}-6)$ of the target in three independent runs ($p_{\mathrm{adj}}=0.001$) and beats expert cycle counts; frozen and applied to five unseen datasets it is never significantly worse than the plain network or the expert rule, and on BBBP it significantly beats both.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!