MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641115380 A) filed by Sri Eshwar College Of Engineering on September 24, 2026, for An Adaptive And Explainable Transformer-Based Framework For Intelligent Phishing Detection And Risk Assessment.
Inventor includes P. Ponni.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: The present invention relates to an adaptive and explainable transformer-based framework for intelligent phishing detection and risk assessment. The framework identifies phishing URLs, webpages, electronic communications, and other suspicious digital resources by analysing multiple phishing-related characteristics. A data acquisition module collects URL, domain, webpage, HTML/DOM, textual, visual, hyperlink, and redirection information, which is processed and normalized by a preprocessing module. A multimodal feature extraction module extracts URL, domain, textual, structural, visual, redirection, and contextual features and converts them into machine-processable representations. A transformer-based detection engine applies an attention mechanism to identify relationships among the extracted phishing characteristics and generates a phishing prediction. A risk assessment module generates a phishing risk score and assigns a predefined threat level to the analysed resource. An explainability module identifies the principal features contributing to the detection decision and generates an interpretable explanation. An adaptive learning module uses validated phishing samples, threat intelligence, analyst feedback, and detection feedback to update model parameters, feature weights, and classification thresholds for adapting to emerging phishing patterns. A response module performs risk-based actions including warning, blocking, flagging, quarantining, or generating security alerts. The framework can be integrated with web browsers, email security systems, network gateways, banking, e-commerce, and enterprise cybersecurity platforms, thereby providing adaptive, explainable, multimodal, and automated phishing detection and risk assessment.
Disclaimer: Curated by HT Syndication.
