arXiv：cs.LG· Jiacheng Zheng, Chang Guo, Zixuan Wang, Xinyu Liu, Hao Chen·· 7 小时前AI 评分42
超越 Scaffold Splits：结构前沿评估揭示 ADMET 模型的隐藏失效
Beyond Scaffold Splits: Structural-Frontier Evaluation Reveals Hidden Failures in ADMET Models
AI 导读
研究提出一种无标签的结构前沿划分方法，保留最稀疏、理化性质最偏远的 scaffold 组，在六个公开 ADMET 任务上，其等权重主误差较 70/10/20 scaffold 对照组中位数上升 87.0%、均值上升 130.3%。
正文
Abstract:Molecular property models are commonly evaluated by holding out Bemis-Murcko scaffolds, yet a scaffold identifier is only one notion of chemical unfamiliarity. We introduce a label-free structural-frontier split that reserves the sparsest and most physicochemically remote scaffold groups, and evaluate it on six public experimental or curated ADMET tasks. Against a 70/10/20 scaffold control with identical acyclic grouping, the frontier inflates equally weighted primary error with a taskwise median of 87.0% and a skew-sensitive mean of 130.3% (descriptive task/seed bootstrap interval, 52.1-246.0%). The mean falls to 75.9% once BBB is removed; that endpoint is the one whose score ranking inverts at the frontier. A message-passing graph-network control still shows a large gap (mean 82.8% over four tasks) and does not invert, so a low-capacity head does not explain the effect. We also test Multi-View Frontier Risk Extrapolation (MV-FREX), a count-adjusted tail-risk penalty over four molecular views, and treat it as a falsifiable probe. It changes normalized frontier error by only 0.16% relative to empirical risk minimization for the perceptron head (interval, -0.43-0.84%) and by -1.9% for the graph network; three fixed robust-penalty controls are likewise inconclusive. Against the published Lo-Hi and DataSAIL splitters, the frontier inflates error more on average, though no split is uniformly hardest. An audit of 31,561 marine natural products further shows that OOD status and agreement with legacy ADMET predictions depend on the molecular view, endpoint, and teacher coverage. Split construction and label provenance are important evaluation constraints in their own right, and the tested training penalties do not resolve the frontier failures we observe.
| Comments: | 15 pages, 4 figures, and 4 tables. Version 3 updates figures and tables caption in the main PDF |
| Subjects: | Machine Learning (cs.LG); Quantitative Methods (q-bio.QM) |
| MSC classes: | 68T05, 68T07 |
| ACM classes: | J.3; I.2.6; I.5.2 |
| Cite as: | arXiv:2607.10729 [cs.LG] |
| (or arXiv:2607.10729v4 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.10729 arXiv-issued DOI via DataCite |
Submission history
From: Jiacheng Zheng [view email]
[v1]
Sun, 12 Jul 2026 12:25:50 UTC (1,936 KB)
[v2]
Sun, 2 Aug 2026 07:57:22 UTC (612 KB)
[v3]
Fri, 7 Aug 2026 02:47:40 UTC (471 KB)
[v4]
Tue, 6 Oct 2026 09:28:57 UTC (616 KB)
来源：arXiv：cs.LG · arxiv.org