Google’s deep learning and AI tech enhances malicious docs detection in Gmail service

28Feb - by aiuniverse - 0 - In Deep Learning


With the ease of access to the internet and mobile phones, people are embracing digital life than ever before. Now, users can hail a cab, transfer money, order food and do other stuff with just a few clicks on a handset’s screen, but you need to have an email ID to avail such services.

Unfortunately, there are many naive users who find it hard to detect what’s genuine and not; they trust everything they receive in the email and this is being misused by cyber criminals.

Hackers are coming up with mischievous techniques such as malware-laced documents attached in the mail, warning notice about impending bank account closure, debit/credit card renewal to fool people to their reveal financial credentials.

Taking cognisance of the issue, Google, over the past few years, has steadily scaled up the security for Gmail against phishing scams and curb the spread of malware. 

In late 2019, the search engine giant launched a new scanner to enhance the detection coverage of Office documents that contain malicious scripts and now the latest results suggest the efforts have born fruits.

Google claims malicious document detection on Gmail has improved by more than 150-percent. 

Here’s the tech behind new scanner on Gmail: Under the hood, Google’s scanner uses a distinct TensorFlow deep-learning model trained with TFX (TensorFlow Extended) and a custom document analyzer for each file type. The document analyzers’ primary job is to parse the document, identify common attack patterns, extract macros, deobfuscate (simplify) content, and perform feature extraction.

Also, the scanner runs in parallel with existing detection capabilities and all of this contributes to the final verdict of Google ‘decision engine’ to block a malicious document.

“Combining different scanners is one of the cornerstones of our defense-in-depth approach to help protect users and ensure our detection system is resilient to adversarial attacks. We will continue to actively expand the use of artificial intelligence to protect our users’ inboxes, and to stay ahead of attacks,” Gmail security team said.

Facebook Comments