PEOPLE | May 9, 2017

Investigative activity starts with data

How the SINTESYS platform helps to find information that is crucial in fighting crime

Crime-as-a-service is a real business model on the Internet, particularly in the deep web, or the non-indexed and most hidden depths of the web (such as new sites, private corporate sites, social networking content, etc.). It is in this dark web, which can be accessed via specific protocols guaranteeing anonymity, where abundant illegal material aimed at illegal trafficking, financial support for crime and much more can be found.

Numerous criminal activities occur (or are discussed) in the deep web, even if analysis of the conversations and content contained in the public web also enables useful material to be intercepted for investigations carried out by the police. In practice this means listening to the web to find trails that are unknowingly left by the perpetrators of crime.

The vastness of the content on the Web – states Vito Morreale, head of the Intelligent Systems and Social Software for Security, Enterprise, Transport, Engineering Infrastructure Laboratory – makes it difficult to intercept relevant and interesting material, but above all to join up the dots, as we often say, or look for a logical thread; a history that unites content on different areas of the Internet.

What is missing in order to implement effective investigative activity?

Many of the investigative and intelligence activities undertaken by various European forces of order are carried out manually and the automatic ones are often basic, while what is needed is the analysis of Big Data, intra- and inter-force collaboration, as well as monitoring of the new tools used by criminals, in order to intercept in time the signals enclosed in the vastness of the Internet. In addition to this, tools are needed that have the capacity to extract and analyse information coming from open sources such as social media and web sites, as well as analysing data extracted from the deep web and dark nets.

SINTESYS (Security Intelligence System) is an open source platform that makes all this possible, giving you the ability to intercept relevant information using text mining and data mining techniques, which enable recognition, for example, of the writing style of a person who has left trails on different sources, so you can join the dots precisely and be able to see important links and correlations.

Therefore in addition to discovering relevant information, SINTESYS make it possible to extract and analyse data to generate the knowledge necessary for investigators to detect events and crimes that leave traces on the network and thus have various modes of visualisation to support decision-making.

SINTESYS – already being tested by Scotland Yard and Madrid Police (in the LASIE project) and soon also by the Italian Carabinieri (ROS), Civil Guard and Judicial Police (within the DANTE project) – is designed to be adaptable across all areas of public security: from the prevention of terrorist attacks to the prevention of common crime and micro-criminality such as bullying. It helps, for example, in identifying: threatening phrases and content on forums, blogs and chat; dangerous subjects in multimedia content published on social media; specific impulsive sounds (screams, shots, broken glasses, etc.) in audio files; dangerous objects (weapons, abandoned suitcases, etc.) in images and videos, as well as relationships between suspect individuals.

What tools are used to extract information?

The platform integrates several technologies useful to textual analysis, implemented with various approaches depending on the investigative phase being conducted. However, by means of strategic partnerships, in-depth semantic analysis is also used to help in understanding the overall meaning of fragmented texts even where there are missing terms or words with figurative meanings, or in the analysis of image/video/audio to extract relevant information from multimedia content.

In addition to this, machine learning approaches, sentiment analysis and a new methodology for social contextualisation of textual documents functional in modelling social groups are used to find out, for example, which people are socially or criminally related to each other.

The information analysed by the platform – says Ernesto La Mattina, responsible for the development of SINTESYS – will be able to navigate through intelligent graphical interfaces, adapt itself to the various types of information sources, actual investigative needs and specific investigation pathways followed by analysts, using a special ability to recognise their habits and on the basis of a psychological study of interaction patterns.

SINTESYS is an example of how analysing Big Data can lead to improvement of the Web’s ecosystem and also therefore, people’s quality of life.

Sonia Montegiove