Given the anonymized Netflix Database and Publicly available IMDb ratings.
Using the IMBD ratings as auxiliary information, and using a weighted scoring algorithm
A user's details can be de-anonymized @ hw1
hw3 Experimenting on different AdblockRules, to find which domains/sites will be blocked. Given a current site, all http requests are analysed and the blocked third party sites are reported.
Third party domains for html objects are created and connectivity graph representing domains from which html objects are loaded
TOR relays and maximum bandwidth-contributing relays are found for a sample request.
hw4 federated learning: created federated learning architecture: This architecture collaboratively learns a shared prediction model while keeping all the training data on central device, created virtual worker nodes that downloads the current model, improves it by learning from data on the node, and then summarizes the changes as a small focused update The main idea is to decouple the ability to do machine learning from the need to store the data in the cloud.
Model inversion attack: given general information like age, height , weight race and type of enzyme, find what dosage is recommended for a particular person
hw5 Used text rank and did text mining on notice of what information a website collects from children
Analyzing user awareness and deployable security solutions to prevent user data abuse through Android Permissions