MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202611051873 A) filed by Kulvinder Singh; and Sanjeev Dhawan on April 23, 2026, for Recommendation System Using Adaptive Moment Estimation (adam) For Better Personalization.
Inventors include Kulvinder Singh; and Sanjeev Dhawan.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: A recommendation system that enhances personalization precision and training stability by integrating Adaptive Moment Estimation (Adam) into model update and optimization processes. The system integrates collaborative filtering, content-based features, and contextual signals. It has a multi-stage architecture with ingestion and preprocessing of data, feature engineering, candidate generation, ranking, and online learning modules. The Adam optimizer is implemented at one or more of the following stages: global model learning, fine-tuning per-user/per-item, and online incremental updates with adaptive learning-rate schedules, moment estimations, and regularization in order to solve data sparsity, non-stationarity, and cold-start issues. Privacy preservation mechanisms, efficient model serving, and integration of real- time feedback are enabled by the system.
Disclaimer: Curated by HT Syndication.
