Gmail thanks to artificial intelligence now blocks 100 million more spam messages every day

The spam mails are the most annoying to manage, but based on the provider to which we entrust our email, these are automatically administered in a special section to allow us to continue with our workflow without annoying interruptions.

Gmail is one of the most used applications in the e-mail even considering that anyone has a Google account, with which we often subscribe to sites that send advertising, or our email ends up in the hands of unwanted addresses. The application provided by Google is one of the best to sort these emails and continues to improve.

Thanks to TensorFlow, a machine learning framework created by Google, Gmail now manages to block 100 million more spam messages every day. In the context of over 1 billion Gmail users, this is not necessarily a huge gain: it works like an additional spam email for 10 users; despite this Google claims that 99.9% of spam e-mails have been blocked and so it is difficult to cover even 0.1%.

"At the scale where we operate, another 100 million are not easy to find, [...] getting the last 0.1% is getting harder, but TensorFlow was great to fill this gap," Neil Kumaran, product manager of Counter Abuse Technology by Google.

Algorithms trained in this way balance a huge number of metrics, from formatting an e-mail to the time of day it was sent. TensorFlow, says Kumaran, simplifies the management of this large-scale data, while the open-source nature of the framework enables rapid integration of new research from the community.

Google says that TensorFlow's integration into Gmail will also allow us to better customize spam filters, but "there is no definition of spam out there," Kumaran adds. despite this, artificial intelligence will help the user to define the best use for himself.