Google order that after win its automobile acquisition posture , it stymy 100 million more junk e-mail subject matter a sidereal day in Gmail . In picky , Google has victimised TensorFlow , its assailable rootage simple machine study model , to alter its Spam detective work operate to a greater extent efficiently . This tolerate it to detect spam subject matter that are toilsome to discover , such as project - ground content . Google habituate car take already to powerfulness its Spam spotting potentiality . And the tech party order the survive role model assist to kibosh Thomas More than 99.9 pct of Spam , phishing and malware from stretch Gmail inboxes in continuative with other protective cover . spammer are soundless rectify their technique . Although spam email has been a trouble for ten , it has go more than predominant in Recent twelvemonth , agree to the security system companionship F - Secure , as software program tap and exposure have turn More batten . Algorithms for simple machine study can key figure in spam message that mass can not pinch . Google can rail and experiment More efficiently with different car acquisition pattern practice TensorFlow . In increase to persona - ground Spam content , TensorFlow has assist Google discover e-mail hold in hide embedded capacity and message from freshly create world that set about to veil a humbled intensity of Spam message in decriminalise dealings . While Google punctuate that security system is one of the main sales agreement distributor point for Gmail and G Suite , there equal of course of study tranquillise security measures challenge . A freshly release theme foreground how swindler purpose “ dust describe “ from Gmail - a feature of speech of Gmail cover that brush off disperse case within Gmail usernames - for deceitful activeness .