arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.18084cs.LG
收藏

平衡图流匹配中的对称性与效率

Balancing Symmetry and Efficiency in Graph Flow Matching

LTS4, EPFL, Lausanne, Switzerland(LTS4,瑞士洛桑联邦理工学院,拉夫斯堡)

Benjamin Honoré, Alba Carballo-Castro, Yiming Qin, Pascal Frossard

AI总结:

通过可控对称调制方案,研究图生成模型中严格等变性带来的计算成本与收敛速度之间的权衡,发现适当调节对称性可在加速收敛的同时避免过拟合。

摘要:

等变性是图生成模型的核心,因为它确保模型尊重图的置换对称性。然而,严格的等变性由于增加了架构约束而提高了计算成本,并且由于模型必须在大量可能的节点置换空间上保持一致而可能减慢收敛速度。我们研究了图生成模型中的这种权衡。具体来说,我们从等变离散流匹配模型出发,在训练过程中通过基于正弦位置编码和节点置换的可控对称调制方案来放松其等变性。实验首先表明,对称性破缺可以通过提供更简单的学习信号来加速早期训练,但代价是鼓励捷径解决方案,可能导致过拟合,即模型重复生成训练集的重复图。相反,适当调节对称性信号可以延迟过拟合,同时加速收敛,使模型在基线训练周期的19%内达到更强的性能。

英文摘要:

Equivariance is central to graph generative models, as it ensures the model respects the permutation symmetry of graphs. However, strict equivariance can increase computational cost due to added architectural constraints, and can slow down convergence because the model must be consistent across a large space of possible node permutations. We study this trade-off for graph generative models. Specifically, we start from an equivariant discrete flow-matching model, and relax its equivariance during training via a controllable symmetry modulation scheme based on sinusoidal positional encodings and node permutations. Experiments first show that symmetry-breaking can accelerate early training by providing an easier learning signal, but at the expense of encouraging shortcut solutions that can cause overfitting, where the model repeatedly generates graphs that are duplicates of the training set. On the contrary, properly modulating the symmetry signal can delay overfitting while accelerating convergence, allowing the model to reach stronger performance with $19\%$ of the baseline training epochs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑