EUVIMEDEUVIMED entdecken
Crossref

ARTIFICIAL INTELLIGENCE AND QUANTUM COMPUTING FOR FINANCIAL CRIME DETECTION IN FINANCIAL NETWORKS

Wisdom Sahr Korsor

EPH-International Journal of Applied Science · 2026 · 12 (4)

DOI: 10.69980/9e3qct53 ↗

Kurz gefasst

Fachartikel in der Zeitschrift EPH-International Journal of Applied Science, 2026. Verfasst von Wisdom Sahr Korsor.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Increasing volumes and complexity of transactions, financial account interconnection, more cross-border transactions, fraud techniques continuously shifting, and other issues are creating a more complex transaction landscape that is increasingly hard to define. AI, machine learning, deep learning, graph analytics, and newfound quantum machine learning (QML) have, therefore, become more and more popular as financial fraud detection and financial-crime prevention tools. Moreover, this paper aims to compare classical machine learning methods, deep learning methods, graph/network methods, quantum machine learning with classical machine learning, and hybrid quantum-classical methods for financial-crime detection based on evidence-based synthesis. The study addresses four research questions concerning existing AI-based financial-crime detection methods, comparative performance, evidence for QML and quantum optimization, and technical and operational implementation challenges. The comparative synthesis of recent peer-reviewed research indicates that the classical family of ML is the most practically matured and support vector machines, neural network approach and anomaly detection approach were among the approaches. The data set sizes for which it has been demonstrated as valid are rather limited and, in some cases, very few in number, for low-dimensional feature spaces, experimental setups, and/or problems. As argued here, the most secure short-term architecture would be to utilize classical AI while complementing it with network analytics to complete the truly essential tasks – but complementing the most with quantum and hybrid processors for subproblems that are hard to solve with classical methods. The paper concludes with an agenda of research avenues concerning realistic networks of financial transactions, standardization of metrics for evaluating financial networks, testing over time, explainability, learning in a privacy-preserving way, test drift, as well as benchmarks (classical/quantum) with reproducible results.

Zitieren

APA

Korsor, W. S. (2026). ARTIFICIAL INTELLIGENCE AND QUANTUM COMPUTING FOR FINANCIAL CRIME DETECTION IN FINANCIAL NETWORKS. EPH-International Journal of Applied Science, 12(4). https://doi.org/10.69980/9e3qct53

Vancouver

Korsor WS. ARTIFICIAL INTELLIGENCE AND QUANTUM COMPUTING FOR FINANCIAL CRIME DETECTION IN FINANCIAL NETWORKS. EPH-International Journal of Applied Science. 2026;12(4). doi:10.69980/9e3qct53

BibTeX
@article{Korsor2026,
  author = {Korsor, Wisdom Sahr},
  title = {ARTIFICIAL INTELLIGENCE AND QUANTUM COMPUTING FOR FINANCIAL CRIME DETECTION IN FINANCIAL NETWORKS},
  journal = {EPH-International Journal of Applied Science},
  year = {2026},
  volume = {12},
  number = {4},
  doi = {10.69980/9e3qct53}
}
Kostenlos starten →