Google has now collaborated with its machine learning framework, TensorFlow, to train additional spam filters for use in Gmail. Through new filters that have been in use since last month, the company revealed that Gmail is blocking 100 million spam messages every day.
Taking into account the more than 1 billion users, the number is not that relevant, even so, Google guarantees that Gmail already blocks 99,9% of spam.
"At the scale we're operating at, an additional 100 million isn't easy to come by," Neil Kumaran, Product Manager for Counter Abuse Technology at Google, tells The Verge. "Getting the last bit of incremental spam is getting harder and harder, but TensorFlow has been great at closing that gap."
Google's email service uses AI in addition to rule-based filters. Even though filters can block the most obvious emails, machine learning looks for new patterns that might suggest an email is untrustworthy.
According to Kumaran, TensorFlow makes it easy to manage this data at scale, and the open source nature of the framework means that new community research can be integrated quickly.
TensorFlow was launched by Google in 2015 and has become an important part of the AI business. The framework is free machine learning that allows developers to build artificial intelligence tools for various tasks.
Google also mentions that the integration of TensorFlow with Gmail will allow you to better customize spam filters.
"There's no definition of spam out there," he adds. But AI could help come up with the best definition for you.
Source: The Verge