MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202631027532 A) filed by Narula Institute Of Technology, Kolkata, West Bengal, on March 9, for 'a system and method for neural network-based predictive caching in edge computing devices.'
Inventor(s) include Gopal Pramanik; Pratap Chandra Roy; Alok Nath Pal; Sonali Mondal; Oyendrila Samanta; Dr. Avijit Mondal; Santosh Das; and Amit Sur.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses a system and method for neural network-based predictive caching in edge computing devices configured to proactively predict future content and service requests using deep learning models and to optimize cache utilization through reinforcement learning-based replacement strategies. The system collects historical access data and contextual parameters, processes the data using a neural network prediction engine to generate probabilistic forecasts of future requests, and prefetches high-probability content into local edge storage prior to actual demand. A reinforcement learning optimization module dynamically adjusts cache replacement decisions based on performance metrics including cache hit ratio, latency, bandwidth utilization, and energy consumption. The invention significantly reduces response time, minimizes backhaul traffic, enhances resource efficiency, and provides scalable deployment across distributed edge computing environments including IoT networks and next-generation communication systems."
Disclaimer: Curated by HT Syndication.
