Due to the challenges of the pandemic, 2020 was the year of digital transformation. In a time of social distancing, it’s not surprising that remote operations were the first to be digitized so that monitoring and diagnosing assets could be done from afar, particularly for utilities, manufacturing, transportation, and fossil fuel exploration. Remote service and condition-based maintenance via Internet of Things (IoT), digital twins, wearable devices, and AI services are increasingly critical. With our introduction today of the OCI Anomaly Detection Service, they are also easier to implement than ever.

Remote anomaly detection — that you can access remotely

OCI Anomaly Detection Service is a robust, scalable and user-friendly AI service that watches large volume multivariate time series data and alerts you when something warrants your attention. Authenticated users can access OCI Anomaly Detection Service, which is part of our public cloud offering, via REST API, command-line interface, development kit, or the Oracle Cloud Infrastructure console. This lets you deploy without in-house data science and ML experts.

Machine learning (ML) and statistical algorithms in this new service reveal complex relationships among signals from diverse system components while providing constant asset surveillance. Catching anomalies with early alerts help you avoid costly disruptions such as operational shutdowns, procedural defects, server malfunctions, and fraud. 

You’ll find OCI Anomaly Detection Service under the Analytics & AI menu in the AI Service section of OCI Console. A step-by-step tutorial is available at Oracle LiveLabs.

Patented, fast, scalable, and actionable prognostics

There are four key features to remember about OCI Anomaly Detection Service:

  1. Patented algorithms: Tested and proven for complex use cases over the past decade, the service is continuously augmented with more than 150 Oracle Labs patents.
  2. Intelligent preprocessing: Improves data quality and speeds time-to-detect. In time series data, common problems include signal mismatch, missing values, and quantized signals. These lead to missed or false alarms. OCI Anomaly Detection Service automatically finds and fixes quality issues for optimal accuracy.
  3. Scalable resources: The OCI Anomaly Detection Services scales responsively depending on training and detection needs, so that developers can focus their effort on solving business problems, not paying scary bills.
  4. Actionable results: Generates a ML model-based estimated value and anomaly score to assess relative significance of anomalies among signals and over time. These let you gauge the severity of incidents and automate business workflows to address them.

Industry use cases for OCI Anomaly Detection Service

Built over a decades-long history of safety-critical applications, OCI Anomaly Detection Service is designed for complex systems with interrelated signals, such as for nuclear power plants, manufacturing, aviation and transportation. Customer use cases include:

  • Utilities: Finding fraud such as power diversion or meter tampering, detecting energy surges, remotely monitoring transformers, and developing condition-based power management systems.
  • Manufacturing: Transforming dashboard-based operational metrics into automated diagnostics that identify anomalies in productivity, factory uptime and product quality to help avoid disruptions.
  • Transportation: Watching fleet operations and cargo condition so that owners and  shippers gain business value.
  • Remote assets: Monitoring ATM machines, cell towers, and building HVAC systems in real-time for optimal customer experience.

This new service is a collaboration between the engineers on the Oracle Labs team and the business experts in many industry verticals. It’s also a collaboration with you. We look forward to seeing how you’ll apply this new service to your own business needs — and as always, we’ll share your stories and continue to grow.