MUMBAI, India, April 17 -- Intellectual Property India has published a patent application (202641043791 A) filed by Sr University, Warangal, Telangana, on April 6, for 'a system and method for multimodal unified representation learning for early prediction of neurodegenerative disease progression.'
Inventor(s) include Chenchu Swetha; and Dr. Chiranjeevi Manike.
The application for the patent was published on April 17, under issue no. 16/2026.
According to the abstract released by the Intellectual Property India: "Based on the multimodal unified representation learning approach, the current invention describes a system and method for the early prediction of neurodegenerative disease progression. The suggested approach uses modality-specific deep learning encoders to integrate neuroimages, genetic data, and EHRs. The variational autoencoder uses a cross-modal attention mechanism to encode the data into a single latent space. The disease progression probabilities and auxiliary cognitive predictions are predicted by the hierarchical prediction module. The relevance gating component's attention mechanism allows the suggested method to handle missing multimodal data and produce interpretable predictions. When compared to unimodal and conventional multimodal approaches, the suggested method's experimental evaluation demonstrates increased prediction accuracy. Keywords Multimodal Learning, Neurodegenerative Disease Prediction, Variational Autoencoder, Cross-Modal Attention, Electronic Health Records (EHR), Disease Progression Modeling."
Disclaimer: Curated by HT Syndication.
