MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541122182 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on Dec. 4, 2025, for 'resource-aware cross-modal knowledge distillation for edge ai deployment.'
Inventor(s) include Dr. Manoov Rajapandy; and Mr. Aaron Mano Cherian.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The present disclosure provides a method for resource-aware cross-modal knowledge distillation in edge artificial intelligence systems. The method includes establishing a distributed computing architecture with multiple computing tiers, monitoring computational resources, and dynamically allocating processing workloads based on resource availability. The method further includes training a teacher model to process multi-modal input data, applying compression techniques specific to different data modalities, transferring knowledge from the teacher model to student models through cross-modal knowledge distillation, and deploying the student models on resource-constrained devices. The distributed computing architecture comprises a three-tier architecture including a cloud layer hosting the teacher model, a fog layer providing intermediate processing capabilities, and an edge layer comprising resource-constrained devices. The compression techniques include structured pruning and quantization for visual data, frequency-domain compression and temporal pooling for audio data, and statistical aggregation with dimensionality reduction for sensor data."
Disclaimer: Curated by HT Syndication.
