What it is

 

CRex is our patented software platform we use to build and deploy AI-driven situational management solutions that monitor, simulate, optimize, and control operations in real time.  It merges the cognitive reasoning power of artificial intelligence, the control of industrial automation, and the performance of Big Data and cloud computing environments.

 

CRex was previously named Big Reasoning Engine.  

 

Semantics, agents and orchestration 

 

SmartCloud's CRex platform is built on three technologies that work together seamlessly: 

• Semantic knowledge base

• AI reasoning agents

• Logic flow orchestration  

 

Our Semantic Knowledge Base delivers real-time data integration, storage , and insight

 

CRex’s semantic knowledge base is a next-generation AI-based database that solves the problem of integrating, storing, and understanding disparate real-time data needed for decision making and control, particularly in Internet of Things (IoT) systems.  

 

Built on OWL2 semantic web standards, our knowledge base models data properties and relationships for on-the-fly, real-time interoperability between many different data sources. There is no need to use separate transformation tools or create completely new structures using data warehouses or data marts.  

 

With semantics, a wide range of complex relationships needed for decision making can be quickly and flexibly modeled. Using a vocabulary that includes physical, network, social, geospatial, causal, and systems connections, our semantic knowledge base enables insight that is meaningful to non-programmers. 

 

To synchronize data in real time while accounting for data latencies, all data is written to persistent storage with two time stamps using patent-pending techniques. Those same techniques support simulations, application rollbacks, audits, and data history snapshot analyses.  

 

AI Reasoning Agents think and act 

 

AI reasoning agents collaborate to decide and act in real time. Compared to traditional centralized or monolithic software, these agents offer an intuitive and powerful approach to modeling the many individual and collective reasoning processes that are part of big, complex, mission-critical operations, such as those found in industrial, government, and commerce settings. 

 

CRex’s AI agents combine reasoning techniques such as heuristic rules, semantic patterns, complex event processing, machine learning using neural networks, probabilistic Bayesian networks, and statistical analysis.

 

Agents scale up to handle massive data loads by intelligently distributing the processing of reasoning logic, whether on field devices or across cloud-based virtual machines.

 

Unlike monolithic software systems, agents can isolate bad data and other problems to a single, autonomous agent, greatly improving application robustness. Agents are also self-healing. For example, a supervisor agent can restart an agent that has timed out due to bad data.  

 

Logic Flow Orchestration coordinates reasoning among internal and external sources

 

Even when faced with massive data loads, CRex’s logic flow orchestration coordinates all application tasks in real time to ensure timely results. It works in combination with CRex's AI agents to account for task concurrencies, timeouts, priorities, communication failures, data latencies, and more. It also executes workflow models that coordinate the processing of calculations, operational models, simulations, reasoning logic, and actions, even with external sources.

 

Stream processing applies reasoning logic while data is in memory, before it is stored, to improve CRex’s ability to analyze big data sets in real time.  

 

Logic flow orchestration’s open, standards-based interfaces employ a publish/subscribe model to simplify real-time integration with disparate data sources, third-party systems, and user interfaces running on desktops, mobile devices, or within other systems. These open interfaces, combined with semantic modeling, fuse data and systems in real time. These elements include structured and unstructured relational database data, non-relational data structures, control and network systems, web interfaces, messaging systems, text messages, and more.

 

To enable offline and online data mining, a data frames interface supports interactive integration with third-party analytics tools, such as spreadsheets and statistical computing using the open source R language.