Real-Time Dynamic Graph Unified Learning Framework for Financial Transaction Risk Management

ZHOU Jun; CAO Yue-tian; HU Bin-bin; ZHANG Zhi-qiang; CHEN Chao-chao

doi:10.12263/DZXB.20220812

您当前的位置：

首页 >

文章列表页 >

Real-Time Dynamic Graph Unified Learning Framework for Financial Transaction Risk Management

PAPERS | 更新时间：2025-12-08

- Real-Time Dynamic Graph Unified Learning Framework for Financial Transaction Risk Management
- ACTA ELECTRONICA SINICA Vol. 51, Issue 10, Pages: 2801-2811(2023)
- 作者机构：
  
  1.浙江大学计算机科学与技术学院，浙江杭州 310027
  2.蚂蚁科技集团股份有限公司机器智能部，浙江杭州 310000
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(72192823;62172362)
- DOI：10.12263/DZXB.20220812
  CLC： TP183;
- Received：24 April 2022，
  
  Revised：2022-08-08，
  
  Published：25 October 2023
- 稿件说明：
移动端阅览
周俊,曹月恬,胡斌斌等.基于实时动态图联合学习框架的金融交易风控技术[J].电子学报,2023,51(10):2801-2811.

ZHOU Jun,CAO Yue-tian,HU Bin-bin,et al.Real-Time Dynamic Graph Unified Learning Framework for Financial Transaction Risk Management[J].ACTA ELECTRONICA SINICA,2023,51(10):2801-2811.
周俊,曹月恬,胡斌斌等.基于实时动态图联合学习框架的金融交易风控技术[J].电子学报,2023,51(10):2801-2811. DOI： 10.12263/DZXB.20220812.

ZHOU Jun,CAO Yue-tian,HU Bin-bin,et al.Real-Time Dynamic Graph Unified Learning Framework for Financial Transaction Risk Management[J].ACTA ELECTRONICA SINICA,2023,51(10):2801-2811. DOI： 10.12263/DZXB.20220812.

摘要

金融交易风险防控是金融风控平台最重要的能力之一.近年来，随着金融风控平台智能化需求的不断升级，对其中相关应用算法的性能要求也水涨船高.目前业界已完成了两代针对金融交易行为的表征学习框架的迭代升级.第一代框架引入了金融交易活动参与者自身的历史行为序列，利用序列模型学习其历史行为特征.第二代框架通过一套实时大数据系统对资金流图进行建模，根据业务专家预定义的业务规则计算出需要的实时特征，并将其输入到后续的判别模型中.相比第一代，第二代框架引入了更多实时动态资金流图上的交互信息，因而取得了不错的性能提升.然而，第二代框架在精细化、智能化和时序建模方面仍存在较大不足.为了解决这些问题，本文针对性地设计了第三代框架，该框架通过动态图表征学习算法，从实时资金流图的原始数据中直接进行表征学习，以此规避了第二代框架中的诸多问题.总的来说，本文在时序信息建模和动态图框架层面均进行了创新性设计.在时序信息建模层面，利用了C

GAT模块（连续时间和上下文感知的图注意力神经网络），在动态多变的资金流图上快速地捕捉了高阶的结构化时序状态与信息.在动态图框架层面，开发了一套联合实时动态图表征框架——RULF，该框架可以实时刻画出金融场景中多用户资金行为中存在的特定模式.将金融场景中多角色联合行为和单角色独立行为进行了显式地解耦，并将多个子图模块联合起来学习，通过学习到更精准的行为表征，从而更进一步地提高下游判别模型的性能.本文将以花呗套现交易识别—一个典型的金融交易风控场景为例，介绍该框架在实际业务场景中的设计思想和实现细节.

Abstract

In recent years

with the continuous escalation of demand in the intelligent financial platforms

the performance requirements for these relevant application algorithms in financial scenarios have also risen. At present

two generations of frameworks about financial role representation learning have been widely used in the industry. The first-generation framework introduced the unique historical sequence of financial roles

and used the sequence model to learn the historical behavior of the role. The second-generation framework put more emphasis on the interaction between roles

built a real-time dynamic graph system through capital flow

and directly obtained the required real-time features through graph calculation according to predefined business rules

and added them to the follow-up discriminant models. Compared with the first generation

it introduced more interactive information

resulting in a good performance improvement. However

the second-generation framework still has great limitations in terms of timeliness

generalization

and ease of use. In order to solve these problems

we design the third-generation framework which directly builds feature from the original real-time capital flow graph through the dynamic graph learning algorithm

avoiding many problems in the second generation. This paper mainly carries on the innovative design in temporal modeling and frame design. In terms of temporal modeling

we design the C

GAT to flexibly capture high-order structured temporal information on dynamic graphs. In terms of framework modeling

we design a real-time dynamic graph framework—RULF

which can better capture and characterize the specific patterns existing in capital behavior in real time financial scenarios. We explicitly separate multi-role joint behavior and single-role independent behavior in financial scenarios

and jointly learn multiple subgraph modules to obtain accurate user representation and performance improvement. A typical interactive financial scenario will be used as a credit cashback example in this article to introduce our design ideas and implementation details in actual business scenarios.

关键词

Keywords

references

YANG S , ZHANG Z Q , ZHOU J , et al . Financial risk analysis for SMEs with graph-based supply chain mining [C]// Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence . New York : ACM , 2021 : 4661 - 4667 .

CAO S S , YANG X X , CHEN C , et al . TitAnt [J]. Proceedings of the VLDB Endowment , 2019 , 12 ( 12 ): 2082 - 2093 .

ZHU X Q , AO X , QIN Z D , et al . Intelligent financial fraud detection practices in post-pandemic era [J]. Innovation (Cambridge) , 2021 , 2 ( 4 ): 100176 .

ZHU Y , LI H , LIAO Y K , et al . What to do next: Modeling user behaviors by time-LSTM [C]// Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence . Melbourne : IJCAI , 2017 : 3602 - 3608 .

KUMAR S , ZHANG X K , LESKOVEC J . Predicting dynamic embedding trajectory in temporal interaction networks [C]// Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . New York : ACM , 2019 : 1269 - 1278 .

TAHEREH P , KOK-LEONG O , BOOI H KAM , et al . Fraud detection: A systematic literature review of graph-based anomaly detection approaches [J]. Decision Support Systems , 2020 , 133 : 113303 .

RAWAT W , WANG Z H . Deep convolutional neural networks for image classification: A comprehensive review [J]. Neural Computation , 2017 , 29 ( 9 ): 2352 - 2449 .

KIPF T N , WELLING M . Semi-supervised classification with graph convolutional networks [EB/OL]. ( 2016 )[2022]. https://arxiv.org/abs/1609.02907.pdf https://arxiv.org/abs/1609.02907.pdf .

VELIČKOVIĆ P , CUCURULL G , CASANOVA A , et al . Graph attention networks [EB/OL]. ( 2017 )[2022]. https://arxiv.org/abs/1710.10903.pdf https://arxiv.org/abs/1710.10903.pdf .

LIU Z Q , CHEN C C , LI L F , et al . GeniePath: Graph neural networks with adaptive receptive paths [J]. Proceedings of the AAAI Conference on Artificial Intelligence , 2019 , 33 ( 1 ): 4424 - 4431 .

BO D Y , HU B B , WANG X A , et al . Regularizing graph neural networks via consistency-diversity graph augmentations [J]. Proceedings of the AAAI Conference on Artificial Intelligence , 2022 , 36 ( 4 ): 3913 - 3921 .

TU K , CUI P , WANG D X , et al . Conditional graph attention networks for distilling and refining knowledge graphs in recommendation [C]// Proceedings of the 30th ACM International Conference on Information & Knowledge Management . New York : ACM , 2021 : 1834 - 1843 .

ZHANG M M , WANG X A , ZHU M Q , et al . Robust heterogeneous graph neural networks against adversarial attacks [J]. Proceedings of the AAAI Conference on Artificial Intelligence , 2022 , 36 ( 4 ): 4363 - 4370 .

LIU H R , HU B B , WANG X , et al . Confidence may cheat: Self-training on graph neural networks under distribution shift [EB/OL]. ( 2022 )[2022]. https://arxiv.org/abs/2201.11349.pdf https://arxiv.org/abs/2201.11349.pdf .

YANG S , HU B B , ZHANG Z Q , et al . Inductive link prediction with interactive structure learning on attributed graph [C]// Joint European Conference on Machine Learning and Knowledge Discovery in Databases . Cham : Springer , 2021 : 383 - 398 .

GUO S N , LIN Y F , FENG N , et al . Attention based spatial-temporal graph convolutional networks for traffic flow forecasting [J]. Proceedings of the AAAI Conference on Artificial Intelligence , 2019 , 33 ( 1 ): 922 - 929 .

HOU Y P , HU B B , ZHAO W X , et al . Neural graph matching for pre-training graph neural networks [EB/OL]. ( 2022 )[2022]. https://arxiv.org/abs/2203.01597.pdf https://arxiv.org/abs/2203.01597.pdf .

LIU Z Q , SHEN Y , CHENG X C , et al . Learning representations of inactive users: A cross domain approach with graph neural networks [C]// Proceedings of the 30th ACM International Conference on Information & Knowledge Management . New York : ACM , 2021 : 3278 - 3282 .

YANG M H , CAO S S , HU B B , et al . IntelliTag: an intelligent cloud customer service system based on tag recommendation [C]// 2021 IEEE 37th International Conference on Data Engineering (ICDE) . Piscataway : IEEE , 2021 : 2559 - 2570 .

高留杰 , 赵文 , 张君福 , 等 . G2S: 基于语义块的知识图谱问答语义解析 [J]. 电子学报 , 2021 , 49 ( 6 ): 1132 - 1141 .

GAO L J , ZHAO W , ZHANG J F , et al . G2S: Semantic segment based semantic parsing for question answering over knowledge graph [J]. Acta Electronica Sinica , 2021 , 49 ( 6 ): 1132 - 1141 . (in Chinese)

车向北 , 康文倩 , 邓彬 , 等 . 一种基于图神经网络的SDN路由性能预测模型 [J]. 电子学报 , 2021 , 49 ( 3 ): 484 - 491 .

CHE X B , KANG W Q , DENG B , et al . A prediction model of SDN routing performance based on graph neural network [J]. Acta Electronica Sinica , 2021 , 49 ( 3 ): 484 - 491 . (in Chinese)

ANDERSON T . The Theory and Practice of Online Learning [M]. Athabasca : Athabasca University Press . 2008 .

ZHU Z B , LIU Z Q , JIN G , et al . MixSeq: Connecting macroscopic time series forecasting with microscopic time series data [EB/OL]. ( 2021 )[2022]. https://arxiv.org/abs/2110.14354.pdf https://arxiv.org/abs/2110.14354.pdf .

QIAN H , WU Q T , ZHANG K , et al . Scope-aware re-ranking with gated attention in feed [C]// Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining . New York : ACM , 2022 : 804 - 812 .

ZHANG K , QIAN H , LIU Q , et al . SIFN: A sentiment-aware interactive fusion network for review-based item recommendation [C]// Proceedings of the 30th ACM International Conference on Information & Knowledge Management . New York : ACM , 2021 : 3627 - 3631 .

DU L , WANG Y , SONG G J , et al . Dynamic network embedding: An extended approach for skip-gram based network embedding [C]// Proceedings of the 27th International Joint Conference on Artificial Intelligence . Stockholm : IJCAI , 2018 : 2086 - 2092 .

SINGER U , GUY I , RADINSKY K . Node embedding over temporal graphs [EB/OL]. ( 2019 )[2022]. https://arxiv.org/abs/1903.08889.pdf https://arxiv.org/abs/1903.08889.pdf .

XU D , RUAN C W , KORPEOGLU E , et al . Inductive representation learning on temporal graphs [EB/OL]. ( 2020 )[2022]. https://arxiv.org/abs/2002.07962.pdf https://arxiv.org/abs/2002.07962.pdf .

HUANG H , FANG Z X , WANG X , et al . Motif-preserving temporal network embedding [C]// Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence . Yokohama : IJCAI-20 , 2020 : 1237 - 1243 .

周俊 , 张志强 , 曹月恬 , 等 . 地理位置和时间感知的表示学习框架 [J]. 智能系统学报 , 2021 , 16 ( 5 ): 909 - 917 .

ZHOU J , ZHANG Z Q , CAO Y T , et al . A geography and time aware representation learning framework [J]. CAAI Transactions on Intelligent Systems , 2021 , 16 ( 5 ): 909 - 917 . (in Chinese)

ZHANG D L , HUANG X , LIU Z Q , et al . AGL: A scalable system for industrial-purpose graph machine learning [J]. Proceedings of the VLDB Endowment . 2020 , 13 ( 12 ): 3125 - 3137 .

HAMILTON W L , YING R , LESKOVEC J . Inductive representation learning on large graphs [EB/OL]. ( 2017 )[2022]. https://arxiv.org/abs/1706.02216.pdf https://arxiv.org/abs/1706.02216.pdf .

NGUYEN G H , LEE J B , ROSSI R A , et al . Continuous-time dynamic network embeddings [C]// Proceedings of the The Web Conference 2018 . New York : ACM , 2018 : 969 - 976 .

LU Y F , WANG X , SHI C , et al . Temporal network embedding with micro- and macro-dynamics [C]// Proceedings of the 28th ACM International Conference on Information and Knowledge Management . New York : ACM , 2019 : 469 - 478 .

SEO Y , DEFFERRARD M , VANDERGHEYNST P , et al . Structured sequence modeling with graph convolutional recurrent networks [C]//CHENG L, LEUNG A, OZAWA S. International Conference on Neural Information Processing . Cham : Springer , 2018 : 362 - 373 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

MoGE: Graph Context Enhanced Multi-Task Recommendation Method

Object Detection Based on EIMYOLO for High-Resolution Remote Sensing Images

Single-Image High Dynamic Range Reconstruction Based on Multi-Attention and Perceptual Weighted Learning

Related Author

ZHOU Jun

CHEN Chao-chao

CAO Feng

ZENG Ke-wen

LI De-yu

LUO Xi-zhao

TAO Chong-ben

Related Institution

College of Computer Science， Zhejiang University

School of Information and Technology, Shanxi University

School of Computer Science and Technology, Soochow University

School of Electronics and Information Engineering, Suzhou University of Science and Technology

Faculty of Information Science and Engineering, Ningbo University

⁰