MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641076142 A) filed by Vellore Institute Of Technology on June 19, 2026, for System For Circuit Fault Detection Using Learnable Spike Encoding.
Inventors include Naushad Manzoor Laskar; Raajaprasanna M; Adhavan K; Kailash Prian Rs; Prasanna Kumar R; and Jonish Jack Kenned.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: ABSTRACT SYSTEM FOR CIRCUIT FAULT DETECTION USING LEARNABLE SPIKE ENCODING A system (100) for circuit fault detection comprises a learnable spike encoder (110) configured to receive an analog circuit signal and convert the analog circuit signal into a spike train. The learnable spike encoder (110) comprises a learnable threshold parameter, a learnable gain parameter, and a learnable offset parameter that are jointly trainable via backpropagation. The system (100) further comprises a feature extractor (120) configured to extract feature representations from the spike train. The system (100) also comprises a spiking neural network classifier (130) configured to classify the feature representations into one of a plurality of fault classes. (Fig. 1)
Disclaimer: Curated by HT Syndication.
