Adaptive Uncertainty-Routed Analysis is introduced, a multimodal email threat detection system that analyzes both the content of an email and its embedded URLs and quantifies prediction uncertainty from a URL classifier.
Abstract
Email spam and phishing attacks remain a critical security threat. Adversaries increasingly exploit large language models to craft contextually convincing malicious messages, and existing spam detection systems often struggle to keep pace. Generalization across diverse and evolving attack scenarios is limited, which reduces effectiveness once these systems are deployed in practice. This paper introduces Adaptive Uncertainty-Routed Analysis (AURA), a multimodal email threat detection system that analyzes both the content of an email and its embedded URLs. AURA is built around two layers: the first quantifies prediction uncertainty from a URL classifier, and only ambiguous messages are escalated to a fine-tuned transformer encoder for semantic analysis. The system is evaluated on eight heterogeneous training corpora together with two held-out real-world corpora spanning a decade of adversarial campaigns. AURA reaches a macro F1-score of 0.9858 in-distribution, and on NazPhish-Eval and GuenterTrap-Eval it maintains 0.9502 and 0.9436, respectively, which is evidence of robust generalization under genuine distribution shift.
Spam and phishing email remain a persistent cybersecurity threat and a primary delivery vector for credential theft, malware, and ransomware. Detection accuracies above 99% are routinely reported. Such figures are typically obtained on small, single-source, monolingual corpora in which near-duplicate campaign templates...
In recent years, phishing attacks have grown exponentially in scale, frequency, and sophistication, placing a significant burden on organizations and security personnel. Attackers leverage advanced obfuscation techniques to evade detection systems and ensure their emails reach users’ inboxes. Additionally, attackers di...
Spam communications remain a persistent cybersecurity threat, consuming network bandwidth, reducing productivity, and delivering malicious content such as phishing links and malware. Traditional rule-based and keyword-centric filtering methods struggle against modern spam that employs contextual manipulation and obfusc...
D. A, K. M, Tarun M et al.· 2026 International Conferenc...· 0 citations
Nowadays, email is the world’s primary communication channel, making it a dominant vector for phishing attacks that exploit urgency, authority, and trust. Advances in Large Language Models (LLMs) have intensified this threat by enabling attackers to generate persuasive, context-aware, and multilingual phishing emails a...
Farouk Aziz, Norbert Oláh· Proceedings of the 13th Inte...· 0 citations
Phishing email detection has increasingly benefited from deep neural models and transformer-based language representations. However, many operational security settings still require detectors that are transparent, low-latency, auditable, and deployable without GPU resources. This paper presents an explainable phishing...
T. Onitiju, Swapnoneel Roy, Debajyoti Mukhopadhyay· International Symposium on N...· 0 citations
A multi-model ensemble that combines adversarial training with spam classifiers diverse in architecture and tokenization substantially improves robustness against both perceptible and imperceptible adversarial attacks while maintaining competitive classification accuracy.
M. Salman, Muhammad Islam, Muhammad Ikram et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026