Automated Spatiotemporal Modeling for Real-Time Data-Driven Actionable Insights

Significant increases in industry requirements for network bandwidth are seen year after year. The exponential growth in streaming data is matched by an increase in the use of machine learning and deep learning to glean actionable insights from these data-ideally in real-time. Demand for artificial intelligence (AI) solutions to a variety of computational needs are likely to increase significantly over the coming years and decades. Meanwhile, the capacity of AI and data scientists to meet current requirements with contemporary approaches, which require continual updating and retraining, is falling short of industry demands for automation on dimensions of critical importance, including training speed, accuracy, trustworthiness, and explainability. In this paper we introduce a hybrid AI approach to computational intelligence which features new self-supervised learning mechanisms, a knowledge model engineered to include support for machine generated ontologies, as well as traditional human-generated ontologies, and interfaces to AGI systems such as OpenNARS, AERA, ONA, and OpenCog. Our hybrid AI system is capable of self-supervised learning of machine-generated ontologies from millions of time series, to provide real-time data-driven insights for large-scale deployments including data centers, enterprise networks, and video analytics. Preliminary results across all the use cases we have attempted to date are promising, but more work is needed to fully map out both the approach's benefits and limitations. This Hybrid AI project, and associated data, are expected to be available as open source in April 2023.