MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126192 A) filed by Dr. Jalaiah Saikam; Dr. Potu Bharath; Dr. C. Srisailam; Dr. K. Anuradha; Manohar P; Nemilidinea Anantharami Reddy; Mulkala Saritha; Dr. K Muralidhar Goud; and Dr. M Shiva Rama Krishna on October 22, 2026, for Ultra Fast Quantum Deep Learning Networks For Resource Constrained Iot Edge Nlp.
Inventors include Dr. Jalaiah Saikam; Dr. Potu Bharath; Dr. C. Srisailam; Dr. K. Anuradha; Manohar P; Nemilidinea Anantharami Reddy; Mulkala Saritha; Dr. K Muralidhar Goud; and Dr. M Shiva Rama Krishna.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: An ultra fast quantum deep learning network for natural language processing (NLP) on resource constrained Internet of Things (IoT) edge devices is disclosed. Input text is segmented into sub-word units, mapped to hashed token embeddings, encoded by a lightweight attention-free context encoder (103) and compressed into bounded rotation angles. The angles are encoded into small variational quantum heads (105) having data re-uploading, whose Pauli-Z expectation values feed a compact output head. After noise-aware and quantization-aware training, each quantum head is compiled into a sparse Fourier quantum circuit surrogate (106) of fixed-point trigonometric coefficients, which the edge device evaluates by Chebyshev recurrence and integer multiply-accumulate operations without simulating quantum states or accessing a quantum processor. A confidence-gated offload controller (107) sends only the compact angle vector to an edge quantum server (108) for low-confidence inputs, while a resource manager (110) adapts heads, surrogate terms and thresholds to energy and latency budgets. Figure accompanying the Abstract:
Disclaimer: Curated by HT Syndication.
