Due to the popularity of portable document format (PDF) and the increasing number of vulnerabilities in major PDF viewer applications, malware writers continue to use the format to deliver malware via web downloads, email attachments, and other methods in both targeted and non-targeted attacks. The topic of how to effectively block malicious PDF documents has received a huge amount of research interests in both cybersecurity industry and academia with no sign of slowing down.

An in-depth peek into the VM running in modern web browsers.

So what exactly is an adversarial autoencoder, and why might you want to use one? They are a clever idea leading to a flexible and general framework for a lot of interesting tasks. Take a deeper dive into AAEs in this paper.

Often, we would like to be able to model probability distributions of high-dimensional data points that represent an overall (much lower dimensional) concept. This lets us learn relevant characteristics of the data in question, and also allows us to easily sample from our data distribution.

As the year 2016 began, a ransomware threat appeared that attacked its victims unlike any previous ransomware attack. SamSam, named after the filename of the earliest sample we uncovered, uses a brutally minimalist, manual approach to target and compromise victims.

As mentioned in our previous1 report about VPNFilter malware, the rst stage implant relies on connecting either to one of 12 hardcoded Photobucket URLs or the Toknowall website to fetch an image that had been specially crafted to contain an encoded form of the command-and-control server’s IP address. The stage one sample extracts the address from the image’s EXIF metadata.