Home page

We present a new method of learning control policies that successfully operate under unknown dynamic models. We create such policies by leveraging a large number of training examples that are generated using a physical simulator. Our system is made of two components: a Universal Policy (UP) and a function for Online System Identification (OSI). We describe our control policy as universal because it is trained over a wide array of dynamic models.

With nearly four billion people still lacking access to the internet, efforts to expand internet access are growing rapidly across the world. Cuba remains one of few emerging nations where this access is still affected by historical trade embargoes and restrictions. Since the 2014 announcement of the normalization of relations between Cuba and the U.S., however, internet access in Cuba is increasing. This work is situated during this time of transition to explore the impacts of increasing internet access on individuals and communities living in Havana.

Spatio-temporal data is often displayed using regional aggregation or heatmaps, which are useful for exploring large distributed trends or working to unearth the cause of more localized behavior. For individual users that live and work in the region, however, these representations are inaccessible and difficult to put into practice. We present a new technique for exploring spatio-temporal data as personal routes through a geographic area.