Google to combat spam on Gmail with an artificial neural network
Google to combat spam on Gmail with an artificial neural network
Google has been doing a pretty good job at filtering spam for Gmail, but the company plans to do even more to combat spam on its email service.
According to Sri Harsha Somanchi, Product Manager, Gmail's spam filter will now use an artificial neural network to identify and block the especially sneaking spam. Thanks to improvements in machine learning, the spam filter is able to detect email impersonation - a major source of phishing scams - by figuring out whether the message is actually sent from a known contact. It also tracks your usage patterns and tailor the spam filter accordingly.
On the other end, Google is helping qualified high-volume senders avoid having their important emails (e.g. confirmation codes, order receipts and flight information) classified as spam. Through Gmail Postmaster Tools, qualified companies can have access to information such as delivery errors, spam reports and reputation so that they can "diagnose any hiccups, study best practices, and help Gmail route their messages to the right places".