December 27th, 2022
Hi everyone,
I wanted to give an update on the progress of my PhD research project, which focuses on using machine learning to detect vault apps in the Aptoide 3rd party Android app store. As some of you may know, I've been working on this project since last March and it will be a part of my dissertation.
So far, we have scraped 500 apps from the app store and collected all the text and images from the individual sites. We have also downloaded any freely available apk files associated with the apps and decompiled them into java files. Using regular expressions, we are searching for any permissions data or imported packages in the files that may suggest that a particular app is a vault app.
The data we gather through this process will be used to train a machine learning program to detect these types of apps. The vault apps will serve as positive examples, while the non-vault apps will be negative examples.
Our goal is to use this research to create a piece of software that can help law enforcement locate hidden vault apps on the devices of bad actors, such as foreign agents or drug dealers, and determine how the information is stored (e.g. locally on the device, in an accessible cloud storage service, or securely encrypted locally on the device).
Overall, the project is progressing well and I'm excited to see the results of our machine learning program once it is trained on the data we've collected. I'll be sure to keep you updated on any further developments.
Thanks for reading!