MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611060029 A) filed by Laith H. Alzubaidi; and Dr. Rohit Sharma on May 12, 2026, for System And Method For Real-Time Anomaly Detection In Streaming Data Using Hybrid Deep Learning Architectures.
Inventors include Laith H. Alzubaidi; and Dr. Rohit Sharma.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: A system for real-time anomaly detection in streaming data using hybrid deep learning architectures is disclosed. The system comprises a streaming data acquisition processor configured to continuously receive heterogeneous operational data from distributed sensing devices, telemetry instruments, communication networks, industrial controllers, and transactional infrastructures. A preprocessing and synchronization processor performs timestamp alignment, signal normalization, packet reconstruction, spectral decomposition, and multidimensional feature tensor generation for continuously generated streaming sequences. A convolutional feature extraction processor identifies localized anomaly signatures and transient operational irregularities from multidimensional feature tensors. A temporal dependency processing processor analyzes sequential behavioral transitions and long-duration operational dependencies associated with evolving anomaly conditions. A contextual attention processing processor generates contextual dependency relationships through adaptive attention computation circuitry for identifying high-significance operational transitions distributed across temporal sequences. A behavioral reconstruction processor generates latent operational representations and reconstruction divergence information indicative of abnormal behavioral conditions.
Disclaimer: Curated by HT Syndication.
