Project Prodigy is Biovista’s technology platform and long term initiative to develop next generation Augmented Intelligence solutions for the healthcare industry. Including tools such as Discover Pro and Vizit, Project Prodigy combines capabilities in AI and Machine Learning, Natural Language Processing, graph databases, federated system architectures and collaborative work UIs to create solutions that support discovery and knowledge intensive analytics in commercial environments.
Complex Systems Problems – Asking the right questions
Understanding disease and developing novel therapies is a “complex system” problem. Unlike classification problems such as face recognition and medical diagnosis, and game playing where there is a limited number of pieces and game rules, in complex problems both the parameters (genes, proteins, pathways, organs) and their relationships (protein interactions, physiochemical interactions) are practically limitless.
In systems class problems, incomplete information can significantly reduce the accuracy of any analytics effort. What is essential is to be able to ask the right questions so as to collect the relevant data on which to then run the analytics.
Biovista’s proprietary Discover Pro is an Augmented Intelligence decision support system designed to help subject matter experts and decision makers in identifying all relevant data – whether obvious, known or non-obvious – so that they can then ask pertinent questions and work on meaningful solutions.
Navigating and integrating multiple Correlation Maps
Biological entities and processes such as genes, pathways, organs and cell functions are studied, understood and can be manipulated at different levels. For example genes are mapped in taxonomies as are their gene-gene interactions. Similarly for pathways, diseases and drugs. When human experts study a disease, or develop a new drug, or personalize a treatment they reason freely and instinctively both within and across these maps. Where it gets hard is when each of these maps is so huge that it becomes impossible for any single or even groups of humans to be aware and reason over each of these maps, let alone reason holistically across all of these maps.
Systems class problems require such a holistic “systems reasoning” across multiple biological interaction maps. Discover Pro is designed to support the navigation of these maps, the discovery of non-obvious links and the integration of these into a coherent mechanistic understanding of disease.
Federated Architecture leverages distributed and private data securely
Big Data Analytics requires access to data. But data that is valuable is not always accessible. Discover Pro has been designed with a federated system architecture that aggregates results computed at the “edge” (client installations) into Biovista’s central servers allowing analytics to combine both public and private information thereby increasing coverage and predictive accuracy, and at the same time allowing greater control of the data by its proper owner.