Deep Reinforcement Learning in Action by Alexander Zai.
Alex holds a BS in Chemical Engineering and Biochemistry from the University of California, Los Angeles (UCLA)., where he graduated in 2014. He co-founded Codesmith and led its Data Science and Machine Learning research group. He has also been involved in advancing TensorFlow.js, focusing on enabling browser-level deep learning and computer vision applications. Today Alex is a Machine Learning Engineer at Amazon AI.
Reinforcement Learning (RL) is certainly one of the most intellectually demanding subfields of Machine Learning. Alex takes a very pragmatic approach by writing this book for the developer who wants to understand the “why” engaging with the “how.” He is utilizing PyTorch as the vehicle, providing a clean, Python interface to very complex algorithms.
Humans learn best from feedback. We are encouraged to take actions that lead to positive results. This reinforcement is applied to computer programs allowing them to solve more complex problems. Deep Reinforcement Learning in Action teaches readers how to program AI agents that adapt and improve based on direct feedback from their environment.
Real demands for RL
This is certainly not an introductory Data Science 101 book. Readers will benefit from having working knowledge of Python at an intermediate level. Regarding Math, Basic Calculus and Linear Algebra is needed in order to understand gradients and tensors. Finally, a base knowledge of Neural Networks for backpropagation is required and I find this fascinating.
Deep reinforcement learning AI systems rapidly adapt to new environments, a vast improvement over standard neural networks. A DRL agent learns like people do, taking in raw data such as sensor input and refining its responses and predictions through trial and error. For instance, Alexander is explaining the Bellman Equation as a way of “back-propagating” future rewards to current actions. This helps the reader gain a sense of how a neural network “learns” a strategy over time.
In conclusion, Deep Reinforcement Learning in Action makes the “black box” feel transparent. The fundamental principles covered remain the bedrock of modern AI systems.