for decades, Elia Lombard hunts criminals through the streets of Naples and other areas of Italy. And it takes more than 20 years combining the study of behaviors, characteristics, customs and strategies of scammers, pickpockets, thieves and muggers and serial with the languages of computer programming. So, this inspector is the superior of the Italian Police has been able to develop XLAW, an algorithm able to predict street crime with precise indications about place, time, geographical area and typology of the author and victims. The police are already using this system for a long time in Naples and Prato. This month has also implanted in Venice, where, a few days after its launch, allowed us to stop commission of a crime to a man stealing in a hotel.
Lombardo, 51, says that the key to making the algorithm was a change of paradigm when attempting to understand the mechanisms of the crime. “Before we used statistical systems. Located the areas most affected by a particular type of crime and sent there on patrol. It is the principle that is still used in almost the entire world,” he explains. But the results in the prevention of the offences were not the ones sought. So this agent licensed in Sciences of the Communication was aimed at university courses in criminology, and things were better.
MORE INFORMATIONSo you know the police if your complaint is untrue (and he gets it right 91% of times) Revolution in police work
“I Understood that some crimes, as burglaries, thefts and robberies, are predictable because they are routine, and are committed usually by professionals,” says Lombardo. The inspector explains that criminals of this type use the same strategies in all the sites. “Choose areas where there is more “prey” and escape routes or shelters adequate. I define game reserves”.
in Addition, these offenders “tend to plan their criminal activity according to a time-scheme monthly, like any worker,” added the police. To discover better methods, was also fixed at what moments they choose to act. “In the majority of cases are phases of operation cyclical, as the arrival of a train or the opening and closing hours of the shops”, she explains. As indicated, more than 85% of the crimes that are committed in a certain territory, despite the that is, have these characteristics.
An algorithm based on models criminals
The study of these dynamics, that Lombardo defines “models criminals”, is the basis of the algorithm XLAW. In it data is inserted corresponding to the complaints and police interventions. “Among the crimes that are taking place in a territory, the program points that belong to these models,” he says. The system launches predictive alerts, which are displayed with circles of different colors on a map through a GIS [Geographic Information System]. “Give information as to the hours and the street that it's going to be a theft, the age range and the form of the action of the alleged perpetrator and the type of victim,” he explains.
So the responsibles of the operational units obtain keys to prevent criminal acts provided for in the following two hours and can give concrete indications to the patrols on how to be distributed in a territory. The same agents on the street, can view alerts on a mobile device.
Lombard began to develop the algorithm in 2004. Seven years later, the Police extended its use to all the territory of Naples. “In these years, the crimes of that type have been reduced by 24%. And the number of complaints and arrests in commission of a crime have increased 27%,” he says. In 2017, the system was implemented also in Prato, a city of Tuscany where the percentage of crimes in the past year had grown, according to a counter-trend to the national average.Tempobet
There, keeps the inspector napolitano, the results of its creation were even better. Since it is used, the street crimes have been reduced by 39% and denunciations and arrests have increased by 54%, according to data offered Lombardo. The success of XLAW seems unstoppable. In early November it announced its implementation in Venice. In the early morning of day 16, a call warned that a thief had come to steal at a hotel in Mestre, the land area of the city. A patrol alerted in advance thanks to the algorithm I was waiting at the door before he left.
more and more, technology conquest of space as a companion of the police forces. Different companies market systems designed to increase the ability to predict crimes. Among the most well known is PredPol, a software that generates alerts of geographical of possible new crimes, according to the probability of risk. Another is COMPAS, the algorithm that calculates the probability that a person will re-offend in the same crime they already committed, and more than once considered controversial because of the possibility of that is based on bias or racist is not more effective than the human.
In Spain, the National Police and Civil Guard do not use systems like XLAW, according to sources in the Ministry of the Interior. “The study of the integration of algorithms of this type in these bodies is still at a level of academic work”, added the same sources. On the other hand, the National Police launched this year a pioneering system of artificial intelligence to detect the false allegations. Sources of this body to ensure that you are also using a computer system that is based on statistical data to indicate what crimes are most common in every area.
The inspector Lombardo ensures that the algorithm, at the time of detecting future crimes, it does not use any bias detrimental to the people involved. “The machine does not focus on the subject and on their biometric features, but in the way they operate.” The software can incorporate autonomously, as well as new models of criminals. “When I developed, in Italy not exist the phenomenon of the baby gangs [bandas criminales street formed by individuals are very young, in some minor cases]. By the year 2016, the algorithm pulled that model only”, it says.
In the opinion of this officer, XLAW can also be useful for other types of crimes, such as vandalism or dumping an abusive waste. “We're doing tests, and we want to work also on the topic of gender-based violence and harassment,” he adds. The effectiveness of the system is by convincing the Ministry of the Interior Italian to extend its use throughout the national territory, says Lombardo.
predictive Tools for the local police
The Chair Eurocop, a project of the University Jaume I (UJI) of Castellón and the company Eurocop, has developed a program for the prediction of crime based on a model of computation probabilistic, different from XLAW. The system is in test by the local police of Castellón.
“Our software draws maps of forecast of risk of any crime in certain places of a city and at certain times of the day,” explains Jorge Mateu, director of the chair and faculty of the UJI.
“ Any type of crime is coded”, adds Ramón García, general director of the company Eurocop.
The program “provides for economic variables, demographic, climatic or flow of movement of people in a particular place”, explains Mateu. The professor believes that, despite the fact that the algorithms predictive existing in the world may be based on different criteria, “there is no one better than the other”, because all “contribute to the improvement of decision making”.
the local police of Rivas-Vaciamadrid (Madrid) has started a primary phase of tests with the algorithm of the Chair of Eurocop. But there is not yet a date planned for start-up operations, according to sources at the city Hall of this town. The responsible of the project of the UJI also collaborate with the police forces of Latin american countries such as Colombia.