MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202631031083 A) filed by Asansol Engineering College, Asansol, West Bengal, on March 15, for 'a multi-modal artificial intelligence system for predicting drug-drug interactions via ensemble fusion of graph learning and semantic embeddings.'
Inventor(s) include Mr. Ujjwal Kumar Kamila; Mr. Arjak Majumder; Mr. Sourav Tewary; Mrs. Hrishita Singh; Mrs. Sristi Choudhury; and Mrs. Tanya Kumari.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "A multi-modal artificial intelligence system and method for predicting drug-drug interactions (DDIs) are disclosed. The invention utilizes a hybrid architecture comprising a Graph Convolutional Network (GCN) for modeling drug networks, a semantic embedding module for biochemical relationship mapping, and a BioBERT-based natural language processing (NLP) module for extracting interaction evidence from unstructured literature. An ensemble fusion layer integrates these multi-modal outputs to generate robust interaction predictions and severity classifications. The system addresses the "cold-start" problem for new medications and provides interpretable clinical explanations to support medical decision-making. Experimental results demonstrate that the ensemble approach achieves an F1-score exceeding 85%, significantly outperforming traditional single-source models. The invention is suitable for real-time integration into clinical prescription systems."
Disclaimer: Curated by HT Syndication.
