arXiv：cs.LG· Hyojung Han, Jongmin Kim, Seung-Hun Jeon·· 3 小时前
CGSD：将训练数据过滤中的选择与再分配分离的反事实门控支持分解
Counterfactual Gate-Support Decomposition: Separating Selection from Redistribution in Training-Data Filtering
AI 导读
研究者提出反事实门控支持分解（CGSD），一种四臂、预算匹配的设计，将门控效应拆分为支持内选择、支持间再分配及两者交互作用。在韩英、韩日检索任务中，仅支持内选择分量超出校准零假设带（+29.7 与 +5.6 留出样本，30/30 正重复），韩英显著、韩日边际；模型置信度阈值门控也复现该模式（+27.3 样本，9/9 正）。数学门控中零合格支持使常规设计不可定义，扩展设计仅在狭窄预算带内可构造。
正文
Abstract:A training-data quality gate changes both which examples survive within regions of the input space and how training mass is distributed across those regions, so a gate-versus-random comparison does not isolate these mechanisms. We introduce Counterfactual Gate-Support Decomposition (CGSD), a four-arm, budget-matched design that crosses gate eligibility with support distribution and decomposes the observed gate effect into within-support selection, between-support redistribution, and their interaction. The practical constraint is finite-pool realizability: we characterize mass and token-assembly conditions under which the required arms can be constructed, together with a support-extinction case in which the ordinary design is undefined. On Korean-English and Korean-Japanese retrieval, the mean within-support selection component is the only one that exceeds its calibrated null band (+29.7 and +5.6 holdout items; 30/30 positive replicates), decisively in Korean-English and marginally in Korean-Japanese; redistribution and interaction remain unresolved at the available power. The same pattern holds for a second, unrelated gate family -- a model-confidence threshold rather than a deterministic verifier -- judged against its own null (+27.3 items, 9/9 positive), so the result is not a property of one gate's mechanism. On a mathematics gate, zero-eligibility supports leave the ordinary design undefined at every budget, while the extended design is constructible only inside narrow budget bands that an exhaustive scan locates and that the natural operating point misses -- the assembly condition separating a property of the pool from a limitation of the packer. CGSD therefore returns a decomposition, a constructible-but-underpowered design, or an explicit non-constructibility result, and says which; it prevents support redistribution from being silently interpreted as sample-level gate quality.
| Comments: | 11 pages, 2 tables, 2 figures |
| Subjects: | Machine Learning (cs.LG); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.19343 [cs.LG] |
| (or arXiv:2610.19343v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19343 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jong Min Kim [view email]
[v1]
Wed, 14 Oct 2026 00:36:40 UTC (53 KB)
来源：arXiv：cs.LG · arxiv.org