Please disable your adblock and script blockers to view this page

Gmail Is Catching More Malicious Attachments With Deep Learning


Gmail
Google
RSA
Microsoft Office
Transportation
Condé Nast
My Personal Information Wired
Affiliate Partnerships


Lily Hay Newman
Elie Bursztein

No matching tags

No matching tags

No matching tags


San Francisco

No matching tags

Positivity     36.00%   
   Negativity   64.00%
The New York Times
SOURCE: https://www.wired.com/story/gmail-catching-more-malicious-attachments-deep-learning/
Write a review: Wired
Summary

But this is exactly the type of pattern-recognition problem where deep learning can be helpful.Currently 56 percent of malware threats against Gmail users come from Microsoft Office documents, and 2 percent come from PDFs. In the months that it's been active, the new scanner has increased its daily malicious Office document detection by 10 percent."Ten percent matters," Bursztein told WIRED. He was also surprised to discover how effective the scanner is at analyzing Microsoft Excel documents, a complicated file format that can be difficult to assess.Though a 10 percent detection increase may not sound like a lot, it's a massive improvement at the scale Google is working on, and any gains are productive given that the threat of malicious documents is a real concern around the world. But we always want to improve our detection capabilities and with malicious documents we chose the one where we could make the most impact for our users."When a full-blown hack is just a rogue Word document download away, users will take whatever extra protections they can get.WIRED is where tomorrow is realized.

As said here by Wired