Computers can do amazing things. There aren’t many things, as a matter of fact, that they can’t help with these days. But, help investigators solve crimes? Yep, even that. Computers can arguably help to solve crimes quicker and more efficiently than simply using manpower to put the puzzle pieces together — all at the touch of a button.
Scientists at the University of Middlesex in London began to tinker with crime solving software in 2014 after receiving a $17 million investment into their research by the European Union, and it was originally conceived as a way to provide early warnings of impending criminal activity, including online fraud and global terrorism. In the course of their experimentation, the team developed an Al System that can match the crime solving capabilities of even the best of the best investigators.
Valcri, as it was named, short for Visual Analytics for sense-making in Criminal Intelligence Analysis, scans millions of pieces of data in police reports at the touch of a button, including text, images, and videos, to suggest links that may be worthy further investigation by those manning the searches. Valcri can help determine the how and why a crime may have been committed, which gives law enforcement more free time to work on building their case. In other words, the new system connects the dots between bits and pieces of otherwise fragmented evidence, allowing investigators to more readily put together their puzzle and transition quickly to confronting perpetrators with their evidence.
Without this technology, crime analysts have historically spent hours upon hours manually sifting through records. Valcri saves valuable time by allowing them instead to gain access to all of the needed information at the touch of a button. The system can even help with routine police work that doesn’t necessary involve solving a meaty case. “An analyst can then say whether this is relevant or not and Valcri will adjust the results,” Neesha Kodagoda, an expert in human computer interactions, said. West Midlands Police has been testing Valcri with around 6.5 million anonymised records gathered over the course of three years, as well as with a force in Antwerp, Belgium, and its creators hope to release the system in real time in the near future.
Much of police work including connecting the dots, but these dots are not always obvious. “The hard part is working out which dots need to be connected,” William Wong, who leads the project at Middlesex, explained. Valcri essentially eliminates the toughest part of the work.
By using the latest in machine learning and facial recognition technology, the system can also improve its results based on input from the analysts. Kodagoda claims, “An experienced analyst needs 73 individual searches to gather all of this information, before manually putting it into an easily digestible form. Valcri can do this with a single click.” So, for example, the programming can hone in on similarities between bullet shells collected from several different crime scenes and determine they were all shot from the same gun, which in turn, helps law enforcement to connect the cases to one perpetrator. These types of things are often extremely time consuming when handled manually.