Simon Crosby: Continuous Intelligence with Machine Learning, Digital Twin and Knowledge Graphs

Simon Crosby, CTO at Swim.ai

Continuous Intelligence (CI) integrates historical and real-time analytics to automatically monitor and update various types of systems, including supply chains, telecommunications networks and e-commerce sites. CI encompasses data ingestion, transformation and analytics, as well as operational “triggers” that recommend or initiate specific real-time actions.

CI casts a wider net than traditional analytics because it includes contextual data, for example related to market behavior, weather patterns or social media trends, that help enterprises operate the core systems more intelligently.

In this episode, Kevin Petrie, our VP of Research interviews Simon Crosby, CTO at Swim.ai, a continuous intelligence software startup. A PhD and former lecturer at the University of Cambridge, Crosby previously served as co-founder and/or CTO of XenSource, Citrix Systems and Bromium.

  • As enterprises accelerate their digital transformations and cloud adoption in the wake of the COVID shock, they need to think creatively about ways to improve operational efficiency and intelligence.
  • System operations generate high volumes of machine data whose increasingly ephemeral value demands real-time analysis and action.
  • Machine learning, digital twins and knowledge graphs each help CI architectures analyze and respond to changing relationships between system components. These components might include delivery trucks, cellular phone towers or e-commerce site nodes.
  • CI technologies layer onto existing infrastructures, potentially including cloud platforms, smartphones or data streaming architectures such as Apache Kafka. They help overcome the performance limitations of traditional databases.
  • Neural networks can assist CI by using simple unsupervised ML models to continuously analyze the relationships of digital twins (“web agents” in Swim.ai terminology) that mirror the state of system components in real time. In-memory processing improves the speed and therefore value of CI solutions based on these neural networks. It predicts the immediate future and recommends action based on those predictions.
  • By mapping the relationships between system components, knowledge graphs effectively create a “LinkedIn platform for things,” enabling things to interact with one another and continuously capture their interactions.
Kevin Petrie

Kevin is the VP of Research at Eckerson Group, where he manages the research agenda and writes about topics such as data integration, data observability, machine learning, and cloud data...

More About Kevin Petrie