MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202531096080 A) filed by National Institute Of Technology, Patna, Bihar, on Oct. 6, 2025, for 'a recommendation system and method for modelling user-item interactions.'
Inventor(s) include Rakshita Mall; and Dr. Maheshwari Prasad Singh.
The application for the patent was published on Jan. 23, under issue no. 04/2026.
According to the abstract released by the Intellectual Property India: "The present invention generally relates to a recommendation system for modelling user-item interactions is disclosed, which considers higher-order and multi-facet user relationships. The system is configured to construct a hypergraph from a social recommendation dataset, wherein hyperedges represent triple-social relations among users and items. A Hypergraph Data Presentation (HYPER) model calculates social relation embeddings to update user and item embeddings, thereby generating final feature embeddings. Subsequently, a Dual Attention (DA) model generates and fuses both local and global semantic features, using a Position Attention Module (PAM) and a Relation Attention Module (REA), respectively, to produce combined final embeddings. An optimization and prediction module then utilizes these combined embeddings and a pairwise ranking loss function to predict user preferences and generate personalized recommendations."
Disclaimer: Curated by HT Syndication.
