Google has recruited its in-house machine learning framework, TensorFlow, to help train extra spam filters for Gmail users. With the new filters in place as of last month, the organization claims Gmail is currently blocking an additional 100 million spam messages every day. With regards to Gmail's 1 billion-plus users, this isn't really a tremendous gain — it works out as one additional blocked spam email per 10 clients — however, Google says Gmail already blocks 99.9 percent of spam, so working out what constitutes that last fragment of a percentage is hard. "At the scale we're working at, an extra 100 million isn't easy to get a hold of," Neil Kumaran, product manager of Counter Abuse Technology at Google, reveals to The Verge. "Getting the last bit of incremental spam is progressively hard, [but] TensorFlow has been extraordinary for closing that gap." Gmail has been using AI notwithstanding rule-based filters for quite a long time. While rule-based filters can hinder the most obvious spam, machine learning searches for new patterns that may suggest an email isn't to be trusted. Algorithms trained in this way balance countless metrics, everything from the formatting of an email to the time of the day it's sent. TensorFlow, says Kumaran, makes dealing with this information at a scale easier, while the open-source nature of framework implies new research from the community can be quickly integrated. TensorFlow was rolled out by Google in 2015, and it has turned into an amazingly important part of its AI business. It's a free machine learning framework that enables developers to create AI tools for an enormous scope of tasks. Fans praise its flexibility and ability to scale, and, obviously, it works consistently with Google's other AI services, urging clients to buy computing power from Google just as off-the-shelf vision and speech algorithms. Google says integrating TensorFlow into Gmail will likewise enable it to better customize spam filters. This procedure has been occurring for a considerable length of time, says Kumaran, with Gmail searching for specific signals from users about what they judge to be spam, however, TensorFlow is "transforming those signals into better outcomes." "There's no one definition of spam out there," he includes. In any case, AI could help work out the best definition for you.
Market Research Vision is a leading market research and market intelligence organization, specializing in research, analytics, and advisory services along with providing business insights & market research reports. Serving as a market research firm, we help our clients in finding: what’s next? We believe in finding creative and innovative solutions through customized and syndicated research reports.