Security system that look at past attacks to help deal with future ones are deeply flawed. Here's an alternative approach that uses the latest machine learning tech. It's no secret that there's a constant game of cat and mouse playing out between hackers and security experts, with both sides working their hardest to stay ahead of the other. While hackers and assorted cyberattackers are always on the lookout for new vulnerabilities to exploit, however, unfortunately security systems can be a bit backwards looking in their approach -- relying on digging back in the archives to try and see how future hacks may play out. That's what Antigena, a machine learning security system developed by British cybersecurity startup DarkTrace is trying to change.