发表机构
University of Macau（澳门大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对已冻结的图聚类模型，本文提出选择性超图精化（SHR）方法，利用属性超图补充高阶关系，仅对可靠节点更新聚类分配，在多组评估中实现了显著的宏观增益，且平均仅少量分配发生变化。
AI 中文摘要
现有图聚类方法通常通过优化模型参数和节点表示来提升聚类性能，但对于已训练完成且处于冻结状态的模型，进一步改进其聚类结果的有效手段仍较为有限。本文研究冻结图聚类的后处理方法：在固定检查点后，该过程不使用标签，也不更新模型参数、节点表示或原始图结构，而是利用属性超图补充普通图难以表达的高阶关系，以此精化现有聚类分配。由于全局超图精化既可能带来性能提升，也可能产生错误更新，本文提出选择性超图精化（Selective Hypergraph Refinement, SHR）方法，该方法从超图生成候选残差方向，并利用图结构、节点属性和匹配空证据评估其可靠性，仅对具有足够支持的节点进行更新，否则保留其原始分配。进一步分析表明，节点是否改变聚类由其自身分配差距和精化的方向强度共同决定。在受控通用套件评估中，15个骨干网络-数据集组合中有13个的平均宏观增益为正，1个完全无动作，1个为负；组合均等宏观增益为0.066个百分点（95%自助置信区间为[0.030, 0.107]个百分点），平均仅0.209%的硬分配发生变化。更广泛的15个原生接口组合评估中，宏观增益为0.137个百分点，平均变化率为0.375%。这些结果表明，冻结聚类输出在训练后仍存在有限但可测量的精化空间，该效果在骨干网络-数据集对间存在异质性，且更广泛的覆盖范围也会增加负迁移的风险。
英文摘要
Existing graph-clustering methods typically improve clustering performance by optimizing model parameters and node representations. Effective means of further improving the clustering results of an already trained and frozen model, however, remain limited. We study post-processing for frozen graph clustering. After checkpoint fixation, the procedure uses no labels and updates neither model parameters, node representations, nor the original graph structure. Instead, it exploits an attribute hypergraph to supplement higher-order relations that ordinary graphs cannot readily express, thereby refining existing cluster assignments. Because global hypergraph refinement can yield both performance gains and erroneous updates, we propose Selective Hypergraph Refinement (SHR). The method generates candidate residual directions from the hypergraph and evaluates their reliability using graph structure, node attributes, and matched-null evidence. It updates only nodes with sufficient support and otherwise retains their original assignments. Further analysis shows that whether a node changes cluster is jointly governed by its native assignment gap and the directional strength of the refinement. In a controlled common-suite evaluation, 13 of 15 backbone-dataset cells had a positive mean macro gain, one produced exact no-action, and one was negative. The cell-equal macro gain was 0.066 pp (95% bootstrap CI, [0.030, 0.107] pp), while only 0.209% of hard assignments changed on average. A broader 15-combination native-interface evaluation yielded a macro gain of 0.137 pp at a mean change ratio of 0.375%. These results indicate that frozen clustering outputs retain a limited but measurable refinement space after training. The effect is heterogeneous across backbone-dataset pairs, and broader coverage also increases exposure to negative transfer.
Comments36 pages, 9 figures; includes appendices
