December 18, 2019 - aiWARE Cognitive Capabilities

A new cognitive capability in the Text cognitive class is now available: Anomaly Detection. Leverage this new capability within the aiWARE AI engine ecosystem or to build your own Anomaly Detection engine.

  • New Anomaly Detection VTN standard

  • Broad use case application: banking, healthcare, cyber security, and more

  • Large, time-series data set support

  • Option of Veritone AI ecosystem engine or onboard your own

  • GraphQL API support 

With the ability to highlight statistically significant values among large datasets, the Anomaly Detection cognitive capability solves a large variety of problems including but not limited to: identifying bank fraud, recognizing cancer cells, and warning of malicious cyber security threats. 

A pre-trained anomaly detection cognitive engine is available today for aiWARE users and customers to work with, and with the release of the Anomaly Detection VTN standard, aiWARE empowers developers to easily onboard their own custom Anomaly Detection engines as well. 

How It Works
From a high level, ingest time-series data that gives the engine a list of numbers representing certain events and corresponding timestamps of when they occurred. The engine will process this and return the likelihood that any given event is anomalous. 

To demonstrate this, we have built an application that integrates and analyzes a folder of images, and returns the one the engine identifies as an anomaly. Check out the demo video below.

Learn More
To learn more about how to leverage Anomaly Detection on aiWARE, check out the Building an Anomaly Detection Engine Tutorial on

Have a Question?
Contact us on Slack at 

Did this answer your question?