MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202641019041 A) filed by Dr. M Reji; Ms. Ranjani K; Ms. Shobana D; Dr. V. Samuthira Pandi; Dr. Jeba Johannah J; Dr. Priyadarsini. K; Dr. M. Vinoth; and Ms. T Ranjitha Devi, Kanyakumari, Tamil Nadu, on Feb. 19, for 'machine learning framework for integrated biomedical and agricultural health analytics.'
Inventor(s) include Dr. M Reji; Ms. Ranjani K; Ms. Shobana D; Dr. V. Samuthira Pandi; Dr. Jeba Johannah J; Dr. Priyadarsini. K; Dr. M. Vinoth; and Ms. T Ranjitha Devi.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "This invention presents a machine learning method that allows for the integrated analysis of both human health information and indicators of crop/soil/plant health to help find connections between the two domains as well as aid in more accurate pre-dictive modeling; and support better decision making from both a precision health standpoint, and a precision agriculture standpoint. A multi-modal data ingestion sys-tem that can take in multiple and different types of information (biomedical (e.g. ge-nomics, proteomics, clinical data (associated with patient history), sensor data from monitoring devices (wearable and otherwise), etc.), and agricultural data (e.g. soil composition, crop phenotyping, pest/disease (sensor data (formal name for this type of data)),...) as well as data from the environment around agriculture is created. A feature engineering and alignment system (pipeline) that adjusts how disparate data sources are matched/associated to a common representation in space (this is done through using adversarial networks between domains and cross-modal embeddings to resolve any differences in how data source types are represented). A deep learning architecture that is built on graph neural networks and temporal convolutional transformer models to jointly model the interactions (of both types) between biomedicine and agriculture) such as environmental exposures that affect chronic diseases (or zoonotic transmission routes from animal/bird to plant/pest) or non-zoonotic transmission routes from plant to human). Predictive and prescriptive analytic components that enable the generation of (from the data created) meaningful insights about "real-world" influences and relationships. The insights include but are not limited to, early indication of risk from agricultural exposure for future human disease (risk of certain diseases resulting from exposure to farmed foods); optimiza-tion of crop management practices with human health being considered downstream; and identifying anomalies in the "real world" for human health relative to the data of the patient/work with agriculture. The invention achieves enhanced predictive accuracy (e.g., 12-28% increase in AUC for multi-task learning benchmarks) by combining similar patterns found in envi-ronmental factors, nutritional factors, and exposure factors from siloed systems that develop independent predictive models within either the biomedical or agricultural industries. This invention supports real-time deployment at or near the point of use (edge deployment), collaborative use via federated learning to ensure the privacy of individual participant data, and scalable cloud-based data analytic processing capa-bilities. In addition to providing solutions for the specific industry needs with respect to food security, One Health initiatives, climate-resilient agriculture, and preventive public health within the biomedical-agricultural nexus, this invention addresses mul-tiple challenges facing the global community."
Disclaimer: Curated by HT Syndication.
