Sharing With Privacy Caring
The possibility of sensing the location of people with the use of mobile devices has dragged a lot of attention from scientists in the last 15 years. At the same time, the raising concern about privacy has made the collection of these mobility traces harder to achieve. Finally, good practices in research articles may soon create a “reproducibility paradox”: content might only be published if the used data are also made public but those might contain sensitive information and may never be brought out.
The purpose of this master thesis will be to generate advanced machine learning models based on real mobility traces. These models will be used to generate “fake” records which can mimic its original source hiding the identity of real subjects.
- Experience with high-level programming languages (python, R, Lua)
- Good understanding of Machine Learning, especially DNN/CNN
- Strong analytical skills to interpret and discuss results
Good to have
- Data Analysis background
- Experience with time series, signal processing, neural-network libraries (Keras, TensorFlow, Torch)
If you are interested in this topic and would like to know more, please send me an email with your CV, transcripts and your motivation to work on this project.
Leonardo Tonetto - tonetto at in dot tum dot de