arXiv cs.AI· Yuen Chen, Yulun Wu, Samuel Sharpe, Igor Melnyk, Nam H. Nguyen, Furong Huang, C. Bayan Bruss, Rizal Fathony·· 15 小时前AI 评分26
BRIDGE：统一序列与图建模的端到端架构
Bridging the Divide: End-to-End Sequence-Graph Learning
AI 导读
研究者提出 BRIDGE，一个将序列模型与图模块耦合在单一目标下的端到端架构，让梯度在两部分间流动以学习任务对齐表示。配套的 TOKENXATTN 是 token 级交叉注意力层，可在相邻序列的具体事件间传递消息。在关系预测与欺诈检测两个设定中，BRIDGE 在排序和分类指标上均优于静态图模型、时序图方法及纯序列基线。
正文 · 原文
Abstract:Many real-world prediction tasks, particularly those involving entities such as customers or patients, involve both {sequential} and {relational} data. Each entity maintains its own sequence of events while simultaneously engaging in relationships with others. Existing methods in sequence and graph modeling often overlook one modality in favor of the other. We argue that these two facets should instead be integrated and learned jointly. We introduce BRIDGE, a unified end-to-end architecture that couples a sequence model with a graph module under a single objective, allowing gradients to flow across both components to learn task-aligned representations. To enable fine-grained interaction, we propose TOKENXATTN, a token-level cross-attention layer that facilitates message passing between specific events in neighboring sequences. Across two settings, relationship prediction and fraud detection, BRIDGE consistently outperforms static graph models, temporal graph methods, as well as sequence-only baselines on both ranking and classification metrics.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2510.25126 [cs.LG] |
| (or arXiv:2510.25126v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2510.25126 arXiv-issued DOI via DataCite |
|
| Journal reference: | Learning on Graphs Conference 2026 |
Submission history
From: Yuen Chen [view email]
[v1]
Wed, 29 Oct 2025 03:06:54 UTC (282 KB)
[v2]
Wed, 1 Apr 2026 18:13:57 UTC (401 KB)
[v3]
Mon, 26 Oct 2026 17:17:00 UTC (319 KB)
来源：arXiv cs.AI · arxiv.org