MUMBAI, India, Jan. 8 -- Intellectual Property India has published a patent application (202511104645 A) filed by Lovely Professional University, Phagwara, Punjab, on Oct. 30, 2025, for 'system for incremental and drift-resilient iot applications classification.'
Inventor(s) include Dr. Nitin Bhardwaj; Monica Gulati; Dr. Pavitar Parkash Singh; and Dr. Kailash Chandra Juglan.
The application for the patent was published on Dec. 12, under issue no. 50/2025.
According to the abstract released by the Intellectual Property India: "A system for incremental and drift-resilient IoT applications classification comprises of Convolutional Neural Network 101 (CNN) which extracts discriminative temporal and statistical features from raw or pre-processed IoT network traffic, a Prototypical Network 102 for creating representative class embeddings (prototypes) for each system type, a random forest classifier 103 for performing final robust system classification using CNN embeddings and prototype-based features, a Principal Component Analysis 104 (PCA) for reducing embedding dimensionality for computational efficiency and stability, a synthetic minority over-sampling technique(SMOTE) for balancing class distributions during training to improve performance on underrepresented system classes."
Disclaimer: Curated by HT Syndication.
