The behavior of a complex system is determined by the effects of its components and their interactions. We use analysis tools from the complexity sciences to understand challenging high-dimensional systems. These learnings enable optimization and inform decisions, or actions, taken by humans and/or machines. By focusing on the underlying drivers, more efficient and effective coordination is achieved between information and resource flows.

Glass Box

Glass Box Technology

Black boxes are systems whose internals are unknown or not easily understandable. Given an input, they produce an output, with no explanation why or how. Black boxes, which are prevalent (and dominant) in data science applications, include neural networks, random forests, and support vector machines. In contrast, our glass box technology provides interpretative diagnostics for produced outputs and permit introspection into the internals of a given system.

Competitive advantages

High-level advantages

Our glass box technology has a number of advantages over black boxes. Interpretative diagnostics permit optimization of predictive performance and reduction of computational complexity. Furthermore, the glass box interpretative diagnostics also provide a structured understanding, thereby answering critical questions: which variables are important? how does the system depend upon these important variables? do they participate in high order interactions?  can the system be simplified?



Systems often make use of information and resources of diverse character and place. Multi-source, multi-target, and multi-structured data are common in today’s enterprises. Our core capabilities are complemented by APIs and leverage RESTful web services. These enable convenient integration within enterprise infrastructures and provision advanced capabilities, such as fully integrated cognitive reasoning platforms.