MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641118154 A) filed by Rajeev Gandhi Memorial College Of Engineering & Technology Autonomous on October 01, 2026, for Deep Learning-Based Intelligent System For Real-Time Anomaly Detection And Predictive Decision Making.
Inventors include Dr. C. Ramakristanaiah; Diyyala Bhavya Sree; Dr. B. Ramasubbaiah; Ms. V. Shobhitha; Ms. K. Jaya Sri; and Ms. S. Tahseen Banu.
The application for the patent was published on October 09, 2026, under issue no. 41/2026.
Abstract: A deep learning-based intelligent system for real-time anomaly detection and predictive decision making is disclosed. A data ingestion module receives streaming data of a monitored system in real time. An anomaly detection module, comprising a deep learning model such as an autoencoder or a recurrent, LSTM, or transformer network, detects anomalies as deviations from the learned normal behaviour, including subtle and complex anomalies that simpler means miss. A prediction module carries the detected anomaly forward, predicting from it, together with the data and a learned relation of anomalies to outcomes, the future state, event, or consequence, such as a failure, damage, or loss, to which the anomaly leads, with a lead time sufficient for action. A decision module then makes, in dependence upon the detected anomaly and the predicted consequence, a decision in real time, an anticipatory action to prevent or mitigate the consequence, an alert, or a control action, in advance of the consequence. The modules are integrated in a single real-time pipeline in which data flows continuously from detection to prediction to decision, and are retrainable, so the system foresees and prevents the consequences of anomalies that mere detection and reporting would not.
Disclaimer: Curated by HT Syndication.
