OptOSS AI 7.2.0 - Release

July 31, 2023

We’re excited to announce the latest release of OptOSS AI, version 7.2.0, which brings a host of user experience improvements!

The changes cover areas such as improved cluster processing workflows, new trends analysis features, and more! 

🚀 What's New?

​1. Hyper parameters search (MVP)

Hyper parameters are used in OptOSS AI for configuring clustering sensitivity and achieving the optimal configuration is essential to receive accurate clustering without duplicates, overlaps, and to combat clusterisation drift. The new “Hyper Parameters search” function allows users to automate searching for the most optimal parameters that fit the target domain. This feature can be used by system backend administrators on the initial step of the basic AI Knowledge Pack creation or when clusterisation re-optimisation is required.

2. Redesign anomaly workflow

This is part of our ongoing user workflow redesign which improves the user experience.

We separate the “Anomalies templates” (anomaly patterns identified by AI) overview from the general flow, however it will be further available for AI professionals from the Clusters view via the Context menu:

This makes the general flow shorter and allows users to navigate straight to the list of anomaly clusters.

3. Black ‘Sun’ diagram upgrade

The black ‘sun’-like diagram which shows your whole network in one view widget is now configurable so users can filter what they see. Users can adjust the view according to their preferences and filter based on Vendor, Model & even Location (City).

4. Automatic classification

From this release, the classification of new “unclassified” clusters can be automated if they have a similarity score of over a pre-specified threshold with an existing cluster. Now, saved prior classifications can be applied automatically by setting thresholds along with each classifier. This feature is available from ‘AI > Restore Clusters classification’ menu

5. Historical Trend analysis

In 7.2 we introduce several new widgets to allow users to spot trend deviations in their anomaly rates. If your network all of a sudden starts experiencing a spike in anomalies, or if there are cyclical fluctuations in the rate, these widgets should make it easy to detect such large scale abnormalities.

There are the following new visualisations available in generic and hierarchical OptOSS AI versions.

Anomaly Trend Graph - visualises the trends for the number of anomalies over time. Filters for clusters, tags, devices are available. This allows configuring scope for monitoring and reacting immediately when the number of anomalies in a particular category went out of the normal. Multiple views with different settings of this widget could be configured and monitored concurrently.

Anomaly Severity Trend Graph - visualises trends of anomalies severity over time.  Filters for clusters, tags, devices, and severity are available.

This allows configuring scope for monitoring by severity and reacting immediately when severity exceeds the trend boundaries.

6. Geo-mapping 

New widget Geo-visualisation for anomalies. Visualises anomalies for a selected period on the map of choice. Different maps can be configured during the integration activities by installers.

⭐ Enhancements

​List of other improvements:

  • Sort and Filter clusters by postcode in the ‘CT-Scan’ view
  • Clusters can be given ‘Custom name’ - operators can give meaningful human-friendly naming to the clusters while the original (system) name will be kept
  • Sharding processor component implementation - allows building very high performance and scaled big data solutions

Contact us via email to have a personal 🔥 Live Demo 🔥, to see  the new features for yourself!