Siren Makes AI Call to Drive Real-time Analytics

Siren investigative intelligence

Organizations can execute massively parallel joins across data feeds to establish relationships between different classes of content.

Siren announced that it has added the ability to discover and then analyze data feeds in real-time to its namesake investigative intelligence platform.

At the same time, version 10.3 of Siren adds entity resolution, deep learning-based predictive analytics and alerting, deep learning-based times series anomaly detection, and the ability to join entities across multiple dashboards.

Each of these capabilities is enabled by artificial intelligence (AI) models created using tools such as TensorFlow or Python that are deployed as Docker containers alongside an instance of a search engine based on Elasticsearch, says Dr. Giovanni Tummarello, Co-Founder and Chief Product Officer at Siren.

See also: Can the Enterprise-In-Motion be Autonomic?

That approach enables Siren to add AI functionality in a modular fashion versus relying on embedded machine learning algorithms embedded with a monolithic instance of Elasticsearch, added Tummarello.

Originally developed under a grant provided by Science Foundation Ireland, Tummarello says Siren differs from other analytics platforms in it based on a single relational data model that enables joins to more easily made across data residing either within the platform or in a third-party database such as Oracle, SQL Server or PostgreSQL. Organizations can then set up data models that are all subsets of the primary data model, says Tummarello.

“Each sub-data model works together flawlessly,” says Tummarello.

To make those data model more accessible, a Dashboard 360 capability now makes it simpler to navigate between joins made across different dashboards, which Tummarello says is not a capability provided by the Kibana-based dashboards that come built into Elasticsearch.

Other new capabilities added to Siren 10.3 include visual tools for exploring graphs created on top of the Siren data model, connectors to both local and remote instances of Elasticsearch, a simplified configuration process, and a wizard that automates connections to Neo4J graph databases. In the next release of Siren, Tummarello says the company will also include over 120 connectors to additional data sources.

Thanks to support for data feeds, Tummarello says use cases of Siren can now be expanded to include, for example, analyzing news stories or cybersecurity alerts. Organizations can then execute massively parallel joins across that data to establish relationships between different classes of content as they best see fit, adds Tummarello.

At a time when organizations are making massive investments in Big Data platforms, Siren is focused on making it simpler to turn all that data into actionable intelligence. To that end, Siren is agnostic in terms of the data source being accessed. Rather than requiring organizations to spend most of their time and effort on integrating various open source tools and frameworks to analyze all that data, Siren provides a unified framework that enables organizations to allocate more time and resources to analyze that data in real-time at a time when organizations are moving away from batch-oriented business processes. The challenge those organizations face, however, is even though most of them can aggregate data in real-time, turning all that data into intelligence that can be acted on in real-time often remains an elusive goal.

Leave a Reply

Your email address will not be published. Required fields are marked *