高阶模型是否因高阶原因而获胜？重新思考超图学习中的性能提升

中文摘要 AI 辅助

高阶模型（例如超图神经网络）在超图学习基准测试中通常优于低阶基线，其优势常被归因于它们利用高阶信息的能力。然而，仅凭更好的性能并不能确立这一解释。因此，我们提出疑问：高阶模型是否因高阶原因而获胜？为探究此问题，我们引入了一个受控的性能归因框架，该框架在保留低阶（即成对）信息的同时扰动高阶信息。在跨越三个任务的25个常用超图学习基准上，我们频繁观察到一种有趣的现象：高阶模型最初优于低阶基线，但在扰动后仍保留其大部分优势。这表明，所观察到的大部分优势在没有高阶信息的情况下仍然可以实现。随后，我们研究了这些剩余差距的潜在低阶解释。我们发现，对低阶基线的简单增强，例如更丰富的成对加权、更多步的成对特征传播以及归一化，缩小了剩余的性能差距，从而支持了部分观察到的优势的低阶解释。我们的分析呼吁超图学习社区重新思考性能归因，区分性能提升与其解释，采用更强的低阶基线，并使用能更好测试高阶信息价值的合适基准。

英文摘要

Higher-order models (e.g., hypergraph neural networks) often outperform lower-order baselines on hypergraph learning benchmarks, and their advantages are commonly attributed to their ability to exploit higher-order information. However, better performance alone does not establish this explanation. We therefore ask: Do higher-order models win for higher-order reasons? To investigate this question, we introduce a controlled performance-attribution framework that perturbs higher-order information while preserving the lower-order, i.e., pairwise, information. Across 25 commonly used hypergraph learning benchmarks spanning three tasks, we frequently observe an intriguing pattern: higher-order models originally outperform lower-order baselines, yet retain most of their advantage after perturbation. This suggests that much of the observed advantage remains achievable without the higher-order information. We then investigate potential lower-order explanations for these remaining gaps. We find that simple additions to a lower-order baseline, e.g., richer pairwise weighting, more steps of pairwise feature propagation, and normalization, reduce the remaining performance gaps, supporting lower-order explanations for part of the observed advantage. Our analysis calls for the hypergraph learning community to rethink performance attribution by distinguishing performance gains from their explanations, adopt stronger lower-order baselines, and use suitable benchmarks that better test the value of higher-order information.

发表机构

• KAIST（韩国科学技术院）
• UNSW Sydney（新南威尔士大学）
• University of Oxford（牛津大学）

机构由 AI 辅助整理，请以论文原文为准。