MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202641097850 A) filed by Kandhuru Sahithya on August 13, 2026, for Explainable Artificial Intelligence Framework For Transparent And Trustworthy Cybersecurity Threat Detection.
Inventors include Kandhuru Sahithya; K. Krishna Reddy; Pillala Sathish; Chidurala Ithvika; and Pilla Balakrishna.
The application for the patent was published on August 21, 2026, under issue no. 34/2026.
Abstract: The present disclosure relates to a computer-implemented system and method for Explainable Artificial Intelligence Framework for Transparent and Trustworthy Cybersecurity Threat Detection. The disclosed system includes a Cybersecurity Data Acquisition Module (CADAM) Set up to gather unstructured and structured cybersecurity data from network devices, cloud infrastructures, intrusion detection systems, and more. Security Information and Event Management (SIEM) platforms, firewalls, cyber threat intelligence feeds, and Internet of Things (IoT) environments. A Cybersecurity Data separation, feature extraction, and feature selection. The Preprocessing and Feature Engineering Module consists of data normalization, log parsing, data separation, feature extraction, and feature selection. The analysis evaluates the protocol in terms of feature extraction, behavioural profiling, anomaly filtering, and contextual. Create standardized cybersecurity data using enrichment and threat correlation. A Hybrid Artificial Intelligence Threat Detection Engine uses transformer networks, graph neural networks, and diffusion models. Transformer networks, graph neural networks and diffusion models are used in Artificial Intelligence Threat Detection Engine. networks, deep neural networks, convolutional neural networks, long short-term memory to detect malware, phishing and other forms of spam, networks, attention mechanisms and ensemble learning models are employed. A type of attack that utilizes malware to steal files and data from a computer system and encrypt them to ransom the victim. An attack where an intruder uses malicious software to encipher data and files on a computer system to extort money from the user. Persistent threats (APTs), botnets, credential compromise, privilege escalation, and zero-day. attacks. An Explainable Artificial Intelligence (XAI) Module gives a transparent and SHAP and LIME for interpretation of cybersecurity decisions using visualization of attention, feature importance, and more. Attribution, counterfactual reasoning, and confidence estimation to boost analysts' trust and decision transparency. These explanations produced are then fed into an Intelligent Cybersecurity Decision Support and Automated Response Module for threat prioritization, The team conducts a cyber risk assessment, provides incident response recommendations and engages in forensic analysis and security. policy optimization. A Cybersecurity Monitoring & Continuous Learning Module continually refreshes threat information based on new trends in attacks and vulnerabilities, Analyst feedback, information, and a Cybersecurity Threat Prediction and Risk Management. Assessment Module forecasts the severity of attacks and the organization's cyber risk. A Security Visualization and Explainable Artificial Intelligence Engine create interactive dashboards; A Model Evaluation can provide summary tables. A Model Evaluation can produce summary tables, while attack timelines, forensic reports and explainability summaries can be provided. and Optimization Module validates and optimizes detection performance using comprehensive analysis. evaluation metrics. A Cybersecurity Decision Support and Monitoring (DSM) Interface is presented below. The following is a Cybersecurity Decision Support and Monitoring (DSM) Interface. Alerts, risk scores, explainable predictions, and recommendations for mitigating risks in the area of cybersecurity. Authorised security personnel, and a Continuous Learning Module can be used to incrementally update the AI-driven models incorporate verified outcome of incidents, threat intelligence from the cyber community, and operational feedback. The disclosed invention provides an enhanced cybersecurity threat detection. accuracy, improves transparency and trustworthiness of AI decision making, decreases false alerts, facilitates privacy preserving collaborative learning and offers an. An intelligent, scalable, adaptive, and explainable cybersecurity framework for the enterprise. and critical infrastructure, industrial control systems, networks, cloud computing platforms. The financial institutions, doctors, hospitals, telecommunications companies, and governments are the ones who need to do this. These are the financial institutions, healthcare organisations, telecommunications networks and the governments that need to do this. agencies.
Disclaimer: Curated by HT Syndication.
