DeepMind’s reinforcement learning lecture series is a collection of 13 videos spanning from 45 minutes to a little less than 2 hours each. The lecture series by DeepMind was created in collaboration with University College London (UCL) for providing information about modern reinforcement learning.
Taught by DeepMind researchers and engineers, the reinforcement learning lecture gradually progresses from fundamentals to advanced concepts.
With concepts like Markov Decision Processes, deep reinforcement learning, and sample-based learning algorithms, off-policy learning, multi-step updates, and more, the lectures are highly focused on Master’s students and above, according to DeepMind.
Read more: Top 10 Innovations by Google DeepMind
The lectures include MDPs and dynamic programming, theoretical fundamentals of dynamic programming algorithms, model-free prediction, model-free control, function approximation, and more.
DeepMind’s reinforcement learning lectures are hosted on their Youtube channel and contain slides in the description of each video for assisting learners to easily follow along with the lectures. Alternatively, you can access the learning rescues from their website.
Heavily focused on advancing the reinforcement learning technology, DeepMind has released a series of learning resources and has even created courses (available on Coursera).