The effectiveness of transformer-based models for BEC attack detection
The effectiveness of transformer-based models for BEC attack detection
Business Email Compromise (BEC) attacks are a significant threat to organizations, with attackers using various tactics to acquire sensitive information and cause financial damage to target firms. These attacks are difficult to detect using existing email security systems, as approximately 60% of BEC attacks do not include explicit indicators such as attachments and links. Even state-of-the-art solutions using Natural Language Processing (NLP) rely heavily on such explicit indicators. This study proposes a transformer-based BEC detection method that can capture linguistic properties of emails so that could reduce the reliance on explicit indicators. Our method of combining BERT and BiLSTM offers the advantage of capturing both global context and local interdependence, resulting in a comprehensive and nuanced understanding of email text. In our experiment, the proposed method outperforms the state-of-the-art solutions, achieving a 0.99% accuracy and this highlights the potential of transformer-based models in detecting BEC attacks.
BERT, BiLSTM, Business Email Compromise BEC, Email Security, Feature Engineering, Phishing, Transformer
77-90
Almutairi, Amirah
93ab82cb-5649-45b5-b6a7-a1ce15446354
Kang, BooJoong
cfccdccd-f57f-448e-9f3c-1c51134c48dd
Fadhel, Nawfal
e73b96f2-bf15-40cb-9af5-23c10ea8e319
13 August 2023
Almutairi, Amirah
93ab82cb-5649-45b5-b6a7-a1ce15446354
Kang, BooJoong
cfccdccd-f57f-448e-9f3c-1c51134c48dd
Fadhel, Nawfal
e73b96f2-bf15-40cb-9af5-23c10ea8e319
Almutairi, Amirah, Kang, BooJoong and Fadhel, Nawfal
(2023)
The effectiveness of transformer-based models for BEC attack detection.
Li, Shujun, Manulis, Mark and Miyaji, Atsuko
(eds.)
In Network and System Security: 17th International Conference, NSS 2023, Canterbury, UK, August 14–16, 2023, Proceedings.
vol. LNCS, 13983,
Springer Cham.
.
(doi:10.1007/978-3-031-39828-5_5).
Record type:
Conference or Workshop Item
(Paper)
Abstract
Business Email Compromise (BEC) attacks are a significant threat to organizations, with attackers using various tactics to acquire sensitive information and cause financial damage to target firms. These attacks are difficult to detect using existing email security systems, as approximately 60% of BEC attacks do not include explicit indicators such as attachments and links. Even state-of-the-art solutions using Natural Language Processing (NLP) rely heavily on such explicit indicators. This study proposes a transformer-based BEC detection method that can capture linguistic properties of emails so that could reduce the reliance on explicit indicators. Our method of combining BERT and BiLSTM offers the advantage of capturing both global context and local interdependence, resulting in a comprehensive and nuanced understanding of email text. In our experiment, the proposed method outperforms the state-of-the-art solutions, achieving a 0.99% accuracy and this highlights the potential of transformer-based models in detecting BEC attacks.
This record has no associated files available for download.
More information
e-pub ahead of print date: 12 August 2023
Published date: 13 August 2023
Additional Information:
Funding Information:
Acknowledgment. This work is supported by the National Nature Science Foundation of China (No. 62102429, No. 62072466, No. 62102430, No. 62102440), Natural Science Foundation of Hunan Province, China (Grant No. 2021JJ40688), the NUDT Grants (No. ZK19-38, No. ZK22-50).
Funding Information:
Acknowledgements. This project has received funding from the European Union’s Horizon 2020 Research and Innovation program under the Marie Sk lodowska-Curie INCOGNITO project (Grant Agreement No. 824015), CONCORDIA project (Grant Agreement No. 830927), SPATIAL project (Grant Agreement No. 101021808) and the Cyprus’s Research and Innovation Foundation (Grant Agreement: COMPLEMENTARY/0916/0031). The authors bear the sole responsibility for the content presented in this paper, and any interpretations or conclusions drawn from it do not reflect the official position of the European Union nor the Research Innovation Foundation.
Funding Information:
Acknowledgment. This work was partially supported by JSPS Grant-in-Aid for Scientific Research (C) 23K11103 and NEC C&C Foundation under Grants for Researchers.
Funding Information:
This research was partially supported by the Chinese Scholarship Council.
Funding Information:
Supported by Chinese Scholarship Council.
Funding Information:
The authors would like to thank the Deanship of Scientific Research at Shaqra University and the Saudi Arabian Cultural Bureau in London (SACB) for allowing the research to be undertaken.
Funding Information:
Acknowledgement. This study is supported by the DFG Cluster of Excellence
Funding Information:
Acknowledgment. This work was partially supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Number JP19H04109, JP22H03592, JP23K16882, and a contract of “Research and development on new generation cryptography for secure wireless communication services” among “Research and Development for Expansion of Radio Wave Resources (JPJ000254)”. which was supported by the Ministry of Internal Affairs and Communications, Japan.
Funding Information:
Acknowledgements. This work was supported in part by the National Nature Science Foundation of China under Grant No. 6197226, the Natural Science Foundation of Guangdong Province under Grant No. 2021A1515011153, and the Shenzhen Science and Technology Innovation Commission under Grant No. 20200805142159001, No. JCYJ20220531103401003.
Venue - Dates:
NSS 2023: 17th International Conference on Network and System Security, University of Kent, Canterbury, United Kingdom, 2023-08-14 - 2023-08-16
Keywords:
BERT, BiLSTM, Business Email Compromise BEC, Email Security, Feature Engineering, Phishing, Transformer
Identifiers
Local EPrints ID: 486167
URI: http://eprints.soton.ac.uk/id/eprint/486167
ISSN: 0302-9743
PURE UUID: e1b39643-b404-4182-bf27-c33bc5250c7f
Catalogue record
Date deposited: 12 Jan 2024 17:31
Last modified: 06 Jun 2024 02:10
Export record
Altmetrics
Contributors
Author:
Amirah Almutairi
Author:
BooJoong Kang
Author:
Nawfal Fadhel
Editor:
Shujun Li
Editor:
Mark Manulis
Editor:
Atsuko Miyaji
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics