Search engine giant Google is utilizing its Machine Learning platform, TensorFlow, to assist train further spam filters for Gmail users. After placing the new filters last month, the company claimed that Gmail is now blocking an additional 100 million spam messages every day. According to the reports, as Gmail has more than 1 billion users, this is not essentially a massive gain.
Gmail has been utilizing Artificial Intelligence and rule-based filters as well, for years. As rule-based filters block the most possible spam, Machine Learning explores new patterns which might recommend an email is not to be trusted. Product Manager of Counter Abuse Technology at Google, Neil Kumaran pointed out that TensorFlow has developed managing this data where it scales easier, while the framework’s open-source nature illustrates new research from the community that can be rapidly incorporated. Introduced in 2015 by Google, TensorFlow has become a vital part of its business processes; because it enables developers to build Artificial Intelligence-optimized tools for a broad array of tasks. TensorFlow functions flawlessly along with Google’s other Artificial Intelligence services, fostering users to purchase computing power, off-the-shelf vision as well as speech algorithms from Google.
The company noted that incorporating TensorFlow into Gmail will also enable it to better personalize spam filters. However, this process has been taking place for years, with Gmail looking for certain indications from users regarding what they judge to be spam, but TensorFlow is turning those indications into better outcomes.