The power of AI in Siren
Artificial Intelligence is a broad term. Exactly what it means often depends on the context it is used in, the solution it relates to, its application and a few more factors. However, one thing is undeniable: the primary goal of AI is to automate and simplify processes.
Within the Elasticsearch-based Investigative Intelligence Platform of Siren, various forms of AI are present to support investigators and analysts in performing their data analyses. Siren’s Investigative Intelligence Platform is a platform in which all the necessary analysis and investigation tools for investigators, fraud analysts, cyber security specialists and anyone else involved in detecting crime are present. From the user-friendly Kibana-based (from Elastic) dashboards, Siren's Federate Plugin allows you to connect to, and seamlessly search, over 200 different data sources, such as SQL. After selecting relevant data from large amounts of data, the platform offers the possibility to visualise, analyse and report this data. In the image below, the different components of the Siren Platform are visually depicted.
In this article, we will explain which functionalities within the platform make use of Artificial Intelligence.
Entity Resolution
Using the Entity Resolution functionality, data is automatically searched for records that may refer to the same entity. This could be a person or a company. This functionality indicates when two entities are not the same, but are significantly related. It is possible to simultaneously search for matches and significant relationships in different data sources, languages, schedules, etc. It is also possible to perform real-time monitoring. If an entity comes into view that is similar or related to the existing entities, this will be reported. In addition, there is the option to automatically review the previously obtained intelligence with the information just obtained.
The two images below visually illustrate how Entity Resolution works.
Natural Language Processing (NLP) & Topic Clustering
The Natural Language Processing (NLP) and Topic Clustering functionalities of Siren also make use of Artificial Intelligence. The NLP functionality helps investigators by automatically extracting entities such as people, locations, and organisations from text records, regardless of the language it is in. The extracted entities are then automatically matched with each other for further link visualisation.
Afterwards, the obtained data can be automatically classified into different taxonomies using the NLP functionality, or Topic Clustering. Such a subdivision may be objects, persons or file components. Then analysts and investigators can navigate through the clusters and then zoom in further on relevant clusters.
In the two images below, you can see both the NLP functionality and the Taxonomy filter.
Visual Toolkit
Siren has an extensive number of functionalities within the Visual Toolkit that make use of Artificial Intelligence. There is the ability to automatically search for similarities within images, detect and search for faces in images and use automatic tagging of objects present in photos. The image below gives a good impression of this.
Anomaly detection
It is also possible to use Siren to automatically detect anomalies in time series. This means that if there are values outside the normal time series, these will be automatically filtered or indicated. The software automatically models trends and periodic returns, also known as periodicities. This can be of added value to analysts and investigators when they want to detect exceptional time series in a large amount of data. In order to take advantage of this AI functionality, analysts/investigators do need to select key parameters prior to the survey.
Since this form of AI uses Elastic Platinum nodes from Elasticsearch, it is only available in the Platinum edition of Siren. Below is a screenshot of the functionality.
Outlier Detection
One of Siren's newest AI features is Outlier Detection. This functionality allows investigators to use behavioural indexes of e.g. phone calls, meetings, financial transactions etc. to detect deviant behaviour of an individual. In other words, to check if the behaviours are different from the norm. This functionality can help investigators and analysts prioritise when strong signals are missing from an investigation. Again, this functionality is only available with the Platinum edition of Siren.
Contact us
Are you curious how Siren can help automate your data analysis in your investigation to promote fraud and crime fighting? Please contact us
