Intellectual Property India Publishes Patent Application for 'Self-Adaptive Machine Learning-based Cyber Defense Framework For Secure Iot Ecosystems' Filed by Dr. B. Dhanalaxmi; Dr. V. Sumalatha; Dr. Boddupally Janaiah; Dr. T. Nishitha; Mrs. Kamala Challa; and Mr. Khaza K B Vali Bhasha SK

MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202541126517 A) filed by Dr. B. Dhanalaxmi; Dr. V. Sumalatha; Dr. Boddupally Janaiah; Dr. T. Nishitha; Mrs. Kamala Challa; and Mr. Khaza K B Vali Bhasha SK, Hyderabad, Telangana, on Dec. 14, 2025, for 'self-adaptive machine learning-based cyber defense framework for secure iot ecosystems.'

Inventor(s) include Dr. B. Dhanalaxmi; Dr. V. Sumalatha; Dr. Boddupally Janaiah; Dr. T. Nishitha; Mrs. Kamala Challa; and Mr. Khaza K B Vali Bhasha Sk.

The application for the patent was published on Feb. 6, under issue no. 06/2026.

According to the abstract released by the Intellectual Property India: "A self-adaptive meta-learning-based cyber defense framework for secure Internet of Things (IoT) ecosystems (AMID-IoT) is disclosed. The framework provides intelligent detection, classification, and autonomous mitigation of cyber threats in heterogeneous and resource-constrained IoT networks. A base intrusion detection and security model is initially trained using diverse IoT network traffic and attack datasets to learn generalized threat representations, wherein the base model is optimized using a model-agnostic meta-learning (MAML) approach to enable rapid adaptation. Upon deployment in a target IoT environment, the base model is adaptively fine-tuned using limited, environment-specific data through few-shot learning to generate a context-aware intrusion detection system capable of identifying malicious activities and anomalous behavior in real time with high accuracy. The framework supports edge-level and gateway-level deployment to ensure low-latency monitoring while minimizing computational and energy overhead on IoT devices. Upon detection of a threat, adaptive defense actions including dynamic access control modification, traffic isolation, device quarantine, and security policy enforcement are autonomously triggered. A continuous learning module periodically updates the model using newly observed network data and evolving attack patterns, thereby reducing false positives, maintaining detection accuracy, and ensuring long-term resilience and robustness of the cyber defense framework against emerging threats."

Disclaimer: Curated by HT Syndication.