"From point clouds to graphs" by Miguel Vaz

7:00pm Socializing

7:15pm "From point clouds to graphs" by Miguel Vaz

For some data problems it is helpful to consider data points relatively to other "near by" points.One way of doing this is by creating a similarity matrix containing the distances between each data point, and analysing it using graph theory.Spectral clustering and correlation networks use this principle very effectively.We will discuss them with some concrete examples and, if time permits, look into t-SNE as a further development of these methods.

8:00pm Networking with beer and pizza

This meetup is sponsored by ConSol* Software GmbH who generously provide their rooms as well as beer and pizza.

Miguel's Bio

Currently at d-fine GmbH I work as a risk management consultant for large financial institutions. In a former life, or so it feels like sometimes, I researched learning methods for interacting with a humanoid robot at the Honda Research Institute in Offenbach. I am passionate about (financial) network models, information visualization, robots and all things communication and interaction.