 So, we work at Urecom and I run a research group there. We are looking at 5G and beyond 5G wireless networks and we think there's a lot of potential for machine learning applications in such networks. We think of two major challenges that could be solved with machine learning. One of them is any scenario where we don't have a proper understanding of physical model to be able to come up with a non-machine learning solution and the other type of scenario where we think machine learning can play a role is anything where we don't have the mathematical tools to really write down equations and solve them properly so machine learning can be used there to be able to come up with flexible solutions. We think of several scenarios really. The one that we are looking at these days mostly is the one where we have radio devices on independent, sorry, autonomous devices that carry radio antennas and that have to come up with decisions mostly on their own so these could be antennas that are mounted on mobile devices or mounted on a piece of infrastructure that are not connected to the rest of the network. For example, flying relay that could be used to provide connectivity to ground users for a number of use cases. The main scenario we are looking at right now is the use of a flying relay that uses machine learning to have the intelligence to position itself, to place itself at the right place to provide connectivity to the ground users. This could be useful when you have a disaster you need to recover from so you have lost infrastructure and you want to send some flying relays to provide connectivity to people on the ground or if you have flash crowd events, spot events, we need to follow people when they move around and you don't necessarily have again the fixed infrastructure to provide the connectivity. Those are the main applications we have in mind. Well I think the big challenge in machine learning is to understand the potential and also the requirement in terms of complexity. Could we implement such algorithms in radio devices and I'm talking about mobile devices that have to be of small size and have to cope with limited energy supply. So this is I think the biggest challenge.