These days an antivirus software is necessary if you want to stay safe and make sure no one is remotely messing with your computer and stealing any files, however antivirus softwares have a big problem and that’s the fact that they can only capture malware and viruses which they know about. For unknown or new threats they first have to be updated and that gives attackers a very brief window of opportunity to attack a computer and most of the times that’s all they need. Researchers are working on ways to fix this dilemma and what they propose means that future antivirus softwares may be able to capture new malware without requiring any updates thus closing the brief window.
Many researchers are working on deep learning algorithms which is capable of spotting new malicious code naturally without need an update. The question is that if an artificial neural network can be programmed to recognize faces it’s possible to program it to recognize code that might be potentially malicious.
Deep Instinct, a startup based in Israel, is working on a deep-learning antivirus software which is claimed to spot new malware with a 20 percent higher accuracy rate than the best antivirus suite available right now. These claims haven’t been independently verified but the basic idea isn’t that different from what researchers at Microsoft and Invincea have presented on the benefits of deep learning detection systems for viruses and malware.
However it’s all said to be in early stages at this point in time so we might have to wait for a bit before there’s antivirus software available that no longer needs us to worry about updating it in order to be fully protected against any potential threats to our online safety and security.