English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The use of Artificial Intelligence (AI) to detect phishing emails isprimarily dependent on large-scale centralized datasets, which opens it up to amyriad of privacy, trust, and legal issues. Moreover, organizations are loathedto share emails, given the risk of leakage of commercially sensitiveinformation. So, it is uncommon to obtain sufficient emails to train a globalAI model efficiently. Accordingly, privacy-preserving distributed andcollaborative machine learning, particularly Federated Learning (FL), is adesideratum. Already prevalent in the healthcare sector, questions remainregarding the effectiveness and efficacy of FL-based phishing detection withinthe context of multi-organization collaborations. To the best of our knowledge,the work herein is the first to investigate the use of FL in emailanti-phishing. This paper builds upon a deep neural network model, particularlyRNN and BERT for phishing email detection. It analyzes the FL-entangledlearning performance under various settings, including balanced andasymmetrical data distribution. Our results corroborate comparable performancestatistics of FL in phishing email detection to centralized learning forbalanced datasets, and low organization counts. Moreover, we observe avariation in performance when increasing organizational counts. For a fixedtotal email dataset, the global RNN based model suffers by a 1.8% accuracy dropwhen increasing organizational counts from 2 to 10. In contrast, BERT accuracyrises by 0.6% when going from 2 to 5 organizations. However, if we allowincreasing the overall email dataset with the introduction of new organizationsin the FL framework, the organizational level performance is improved byachieving a faster convergence speed. Besides, FL suffers in its overall globalmodel performance due to highly unstable outputs if the email datasetdistribution is highly asymmetric.

Evaluation of Federated Learning in Phishing Email Detection