MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202541133995 A) filed by Mr. M. Suresh; Mr. G. Suresh Kumar; Mr. P. Jayachandran; Mr. A. T. Barani Vijaya Kumar; Mr. K. Saravanan; Mr. M. Mohamed Nizarudeen; Mr. C. Ramachandran; Ms. M. Vaishnavi; Mrs. Subashri R; and Mrs. R. Rajavaishnavi, Tiruchirapalli, Tamil Nadu, on Dec. 31, 2025, for 'deep neural network architecture for dynamic sequential data prediction using multi-stage attention and temporal feature refinement.'
Inventor(s) include Mr. M. Suresh; Mr. G. Suresh Kumar; Mr. P. Jayachandran; Mr. A. T. Barani Vijaya Kumar; Mr. K. Saravanan; Mr. M. Mohamed Nizarudeen; Mr. C. Ramachandran; Ms. M. Vaishnavi; Mrs. Subashri R; and Mrs. R. Rajavaishnavi.
The application for the patent was published on Jan. 9, under issue no. 02/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses a deep neural network architecture for dynamic sequential data prediction using multi-stage attention and temporal feature refinement. The system processes time-ordered data through hierarchical temporal encoding followed by multiple attention stages that adaptively identify relevant temporal segments across different time scales. Temporal feature refinement modules are interleaved within the architecture to suppress noise, reduce redundancy, and enhance salient temporal patterns, thereby improving robustness and prediction accuracy. The architecture supports adaptability to non-stationary data distributions, efficient handling of long sequences, and interpretability through multi-level attention analysis. By progressively refining temporal representations and reallocating attention based on contextual relevance, the proposed invention enables accurate, stable, and scalable prediction for complex sequential data across diverse application domains."
Disclaimer: Curated by HT Syndication.
