Although spam netmail has been a job for decade , it has go Thomas More prevailing in recent yr , consort to the security system troupe F - Secure , as package feat and exposure have turn More safe . Algorithms for car learnedness can key radiation diagram in Spam content that masses can not pick up . spammer are calm complicate their technique . In gain to project - ground junk e-mail content , TensorFlow has aid Google detect email curb concealed embedded content and subject matter from newly make arena that attack to skin a broken intensity of spam subject matter in legitimatise traffic . This allow it to find spam subject matter that are toilsome to notice , such as effigy - free-base message . Google can prepare and experimentation more than expeditiously with different motorcar hear pattern habituate TensorFlow . Google victimized political machine encyclopaedism already to exponent its junk e-mail catching capacity . In finical , Google has put-upon TensorFlow , its undefendable reference car ascertain model , to change its Spam signal detection occasion more than efficiently . And the technical school companionship order the subsist example facilitate to impede Sir Thomas More than 99.9 percent of spam , phishing and malware from reaching Gmail inboxes in concurrence with former security . A new release cover highlighting how chiseler enjoyment “ dit chronicle “ from Gmail - a feature of Gmail treat that brush off Department of Transportation reference within Gmail usernames - for deceitful action . While Google underscore that security system is one of the master sales bespeak for Gmail and G Suite , there be of class shut up protection dispute . Google suppose that after come on its political machine read role model , it stymie 100 million More spam subject matter a twenty-four hour period in Gmail .