According to Google, Gmail’s current machine learning model helps in blocking 99.9% of threats from reaching users’ inboxes. However, in order to improve its detection capabilities Google added more security against malicious documents in Gmail.
In order to make the model even more accurate, the company has added a new generation of document scanners that rely on deep learning to improve the company’s ability to discover malicious documents. “Strengthening our document detection capabilities is one of our key focus areas, as malicious documents represent 58% of the malware targeting Gmail users. We are still actively developing this technology, and right now, we only use it to scan Office documents” says Google.
The company states that the new scanner uses TensorFlow deep-learning model trained with TFX (TensorFlow Extended) and a customised analyser for each file type. The analysers are responsible for parsing the document, identifying common attack patterns, extracting macros, deobfuscating content and performing feature extraction.