We are witnessing a proliferation of mobile devices integrated with global positioning (GPS) functionality and internet connectivity in recent years. Such technological advancements, along with increased individual mobility is contributing to a notable progress in the development of location-based services (LBS). Popular examples of LBS being UBER, Foursquare, Tinder amongst others. However, sharing sensitive location data with third party entities, exposes users to critical privacy risks. As a result, the success of such applications depends on establishing a rational tradeoff between the user privacy and service utility. In this contemporary impatient modern world, users are demanding services to be pushed ahead of time. In the coming era, we will witness mobility prediction established as a key paradigm in LBS. Think of it, as your waiting time for UBER is reduced as UBER can predict where you will go, or even better, your Tinder date is waiting for you even before you go to a new destination ;).

The talk will focus on current trends in privacy preserving mechanisms for predictive analytics in the context of geospatial mobility. A brief introduction to how a large amount of user spatiotemporal data is handled by telecom providers to preserve user privacy and the resulting vulnerabilities. How MLSP (Machine Learning for Signal Processing) can be applied to address such vulnerabilities and provide conclusive and generalisable results.