Imperial College London > Talks@ee.imperial > Featured talks > Reinforcement Learning for Proactive Content Caching in Wireless Networks

Reinforcement Learning for Proactive Content Caching in Wireless Networks

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Joan P O'Brien.

Next generation wireless networks need to adopt technologies that can handle the ever-growing mobile data traffic and services. Proactive content caching at the wireless network edge can enable future networks to meet these demands by exploiting limited cache memories at wireless access points or user equipments (UEs). In this talk, I will propose proactive content caching directly at mobile UEs, which request contents from a non-stationary content library; new contents are added to the library over time, which remain relevant for a random lifetime. The transmission cost of requested contents (energy or bandwidth) depends on the channel quality at the time of delivery. I will use reinforcement learning (RL) to solve this highly challenging stochastic optimisation problem with prohibitively large state and action spaces. In particular, I will employ policy approximation techniques for compact representation of complex policy structures, and policy gradient RL methods to train the system. The problem complexity further grows in the case of multiple receivers served over a wireless broadcast channel, for which I will propose a multicast-aware proactive caching scheme using deep RL. Bio: Samuel O. Somuyiwa received the B.Sc. degree in Electronic and Electrical Engineering from Obafemi Awolowo University, Nigeria, in 2011, and the M.Sc. degree in Communications and Signal Processing from Imperial College London, UK, in 2014, where he is pursuing the Ph.D. degree. His research interest is in the application of machine learning for optimisation of wireless networks.

This talk is part of the Featured talks series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

Changes to Talks@imperial | Privacy and Publicity