AutoAudit: Mining Accounting and Time-Evolving Graphs

Lee, Meng-Chieh; Zhao, Yue; Wang, Aluna; Liang, Pierre Jinghong; Akoglu, Leman; Tseng, Vincent S.; Faloutsos, Christos

doi:10.1109/bigdata50022.2020.9378346

Cited by 17 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is a lot of work on Anomaly detection (Akoglu, Tong, and Koutra 2015;Liu, Ting, and Zhou 2008;Lee et al 2021); on dense sub-graph detection, which are usually suspicious (Hooi et al 2016;Shin, Eliassi-Rad, and Faloutsos 2016); on unsupervised clustering, that group nearby points and indicates groups and trends in the dataset (Hamerly and Elkan 2003;Ester et al 1996;Ankerst et al 1999;; on (semi-)supervised methods, when only some of the nodes have labels (Ester et al 1996;Ankerst et al 1999;Hamerly and Elkan 2003); on time-evolving graphs (Kazemi et al 2020;Lee et al 2020); on graph visualization (Stolper et al 2014;Chau et al 2011;Zheng et al 2022); and on call graphs (de Melo et al 2010;Akoglu, de Melo, and Faloutsos 2012).…”

Section: Related Workmentioning

confidence: 99%

TgrApp: Anomaly Detection and Visualization of Large-Scale Call Graphs

Cazzolato

Vijayakumar

Zheng

et al. 2023

AAAI

View full text Add to dashboard Cite

Given a million-scale dataset of who-calls-whom data containing imperfect labels, how can we detect existing and new fraud patterns? We propose TgrApp, which extracts carefully designed features and provides visualizations to assist analysts in spotting fraudsters and suspicious behavior. Our TgrApp method has the following properties: (a) Scalable, as it is linear on the input size; and (b) Effective, as it allows natural interaction with human analysts, and is applicable in both supervised and unsupervised settings.

show abstract

Section: Related Workmentioning

confidence: 99%

TgrApp: Anomaly Detection and Visualization of Large-Scale Call Graphs

Cazzolato

Vijayakumar

Zheng

et al. 2023

AAAI

View full text Add to dashboard Cite

show abstract

“…Anomaly detection has been deployed in financial services systems to capture the differentiating characteristics of fraud in the immense quantities of financial transaction data [65], [66]. It provides insights on the data patterns by focusing on the distinct characteristics in terms of connectivity, entity characteristics, flow, traffic patterns (both locally and globally), sub-graph characteristics and events in the graph representations [67,68], [69], [66].…”

Section: Graph Anomaly Detectionmentioning

confidence: 99%

Graph Computing for Financial Crime and Fraud Detection: Trends, Challenges and Outlook

Kurshan

Shen

2020

Int. J. Semantic Computing

View full text Add to dashboard Cite

The rise of digital payments has caused consequential changes in the financial crime landscape. As a result, traditional fraud detection approaches such as rule-based systems have largely become ineffective. Artificial intelligence (AI) and machine learning solutions using graph computing principles have gained significant interest in recent years. Graph-based techniques provide unique solution opportunities for financial crime detection. However, implementing such solutions at industrial-scale in real-time financial transaction processing systems has brought numerous application challenges to light. In this paper, we discuss the implementation difficulties current and next-generation graph solutions face. Furthermore, financial crime and digital payments trends indicate emerging challenges in the continued effectiveness of the detection techniques. We analyze the threat landscape and argue that it provides key insights for developing graph-based solutions.

show abstract

“…Anomaly detection (AD), also known as outlier detection, is a crucial learning task with many real-world applications, including malware detection (Nguyen et al 2019), anti-money laundering (Lee et al 2020), rare-disease detection (Li et al 2018) and so on. Although there are numerous detection algorithms (Aggarwal 2013;Pang et al 2021;Zhao, Rossi, and Akoglu 2021;Liu et al 2022), existing AD methods assume the availability of (partial) labels that are clean (i.e.…”

Section: Introductionmentioning

confidence: 99%

ADMoE: Anomaly Detection with Mixture-of-Experts from Noisy Labels

Zhao

Zheng

Mukherjee

et al. 2023

AAAI

View full text Add to dashboard Cite

Existing works on anomaly detection (AD) rely on clean labels from human annotators that are expensive to acquire in practice. In this work, we propose a method to leverage weak/noisy labels (e.g., risk scores generated by machine rules for detecting malware) that are cheaper to obtain for anomaly detection. Specifically, we propose ADMoE, the first framework for anomaly detection algorithms to learn from noisy labels. In a nutshell, ADMoE leverages mixture-of-experts (MoE) architecture to encourage specialized and scalable learning from multiple noisy sources. It captures the similarities among noisy labels by sharing most model parameters, while encouraging specialization by building "expert" sub-networks. To further juice out the signals from noisy labels, ADMoE uses them as input features to facilitate expert learning. Extensive results on eight datasets (including a proprietary enterprise security dataset) demonstrate the effectiveness of ADMoE, where it brings up to 34% performance improvement over not using it. Also, it outperforms a total of 13 leading baselines with equivalent network parameters and FLOPS. Notably, ADMoE is model-agnostic to enable any neural network-based detection methods to handle noisy labels, where we showcase its results on both multiple-layer perceptron (MLP) and the leading AD method DeepSAD.

show abstract

AutoAudit: Mining Accounting and Time-Evolving Graphs

Cited by 17 publications

References 18 publications

TgrApp: Anomaly Detection and Visualization of Large-Scale Call Graphs

TgrApp: Anomaly Detection and Visualization of Large-Scale Call Graphs

Graph Computing for Financial Crime and Fraud Detection: Trends, Challenges and Outlook

ADMoE: Anomaly Detection with Mixture-of-Experts from Noisy Labels

Contact Info

Product

Resources

About