Presenting and researching during covid: Robert Choudhury

During the lockdown I was researching the implementation of sandboxes with mobile operating systems on the internet. After producing a program to survey these platforms and setting up a reporting server I was able to ascertain that sensor implementations were vulnerable to trivial attacks because of the lack of dynamic values witnessed during the survey. I furthered this work by looking into how researchers would run malware samples on local physical devices which would report the results and then be reset ready for the next execution. This setup has the benefit of including all the sensors of a mobile device and real-world readings.

 I decided to imitate such a setup at home using a physical device attached via USB to a reporting server to see if I could exploit the real-world readings to tell me something about the nature of the app analysis. From this I created an app that contained a reverse Turing test to act as a trigger for malware only when a device is performing a recognised human activity such as walking. It was also possible to invert this model to flag a lack of activity i.e., the device is still. The challenges of research during covid were focused on data collection, this meant that the creation of my own activity recognition model was not possible, so I had to make use of an existing model.

 

I recently presented this work remotely and I was given the option of pre-recording the presentation (this provided resilience if your internet connection failed).  ‘Discord’ was used so that anyone presenting could communicate with the organisers just in case there were last minute changes to the ordering or technical issues. I used the Royal Holloway presentation template for PowerPoint to produce my slides. I chose to present live and fielded questions at the end. All in all, I enjoyed the experience and what I found interesting was the amount of data we had to leave out of the presentation due to RRI.