Arguably the biggest obstacle to the widespread implementation of “Big Data” machine learning (ML) and artificial intelligence (AI) is the massive amounts of data and associated infrastructure these methods require to be predictive. This makes the implementation of “Big Data” ML/AI complicated, costly, unwieldy, and relegated to server based or on cloud, where data storage, management and computational power is sufficient to manage it.

PredictiveIQ™ is developing a new class of data-driven and physics informed ML/AI that utilizes advanced mathematics and/or physics priors to reduce by orders of magnitude (1,000X) the amount of data required to train ML/AI algorithms (i.e. “Small Data”). This in turn enables real-time, which can be done on-platform, on-edge or on-cloud and significantly improves the predictive accuracy of the ML/AI algorithm, thereby making them actionable.