PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
October 18, 2025Open Access

Robust ML-based Detection of Conventional, LLM-Generated, and Adversarial Phishing Emails Using Advanced Text Preprocessing

View Full Paper
Ask AI
Bookmark
Share

Authors

DKDeeksha H KulalCAChidozie Princewill ArannonuAAAfsah Anwar

Discussion

Loading...

Member takes

Overview

This system demonstrates enhanced detection accuracy against phishing emails using machine learning, highlighting resilience against LLM-generated content.

Key Points

  • Our approach achieves a detection accuracy of 94.26%, successfully identifying both conventional and LLM-generated phishing emails.
  • Text preprocessing methods, including spelling correction and word splitting, enhance the model’s ability to detect adversarial modifications.
  • Evaluation against adversarial phishing samples shows the model's robustness against evolving AI threats and detection challenges.
  • Integration of machine learning algorithms and NLP techniques delivers improved performance over traditional phishing detection systems.

Cite This Study

Kulal et al. (2025) studied this question.

synapsesocial.com/papers/68f3793258f37cefb60d3549https://doi.org/10.48550/arxiv.2510.11915
View Full Paper
Ask AI
Bookmark
Share