Reinforcement learning in the real-world

Wanna learn with me? 📜

Reinforcement learning is a powerful framework and a family of algorithms you can use to solve many real-world problems.

Many call it the third big subdiscipline in ML, together with supervised ML and unsupervised ML. It is actually a well-matured field, that recently got a breed of fresh air coming from the ML world thanks to neural networks architectures and the so-called Deep Learning.

The most iconic example at this time is AlphaGo, the first computer program to play the game of Go (probably) better than any other human has ever done.

I always thought that Reinforcement Learning is a hard discipline. Numerous times I skimmed through tutorials, code snippets, and maths formulas to understand how it works and how can I develop real-world solutions using my preferred language, Python 🐍❤️.

And I failed every single time.

The main reason being, excessive complexity. If you start learning RL by reading papers in Deep RL you will fail. Terribly.

Why? Because cutting-edge RL solutions are based on several layers of complexity, including:

  • dynamic programing and optimization.
  • hyperparameter tunning.
  • traditional ML
  • Deep ML.
  • … and a bunch of RL-specific tricks 🤯🤯🤯

You need to cover them, step by step, to REALLY understand how you can apply RL to the problems you care about.

At the end of the day, playing Go is amazing, but what about building RL solutions that make a positive impact on your business by solving real-world problems?

This is why I decided to create an online course on Reinforcement Learning. A hands-on course, 100% free for my subscribers.

Subscribe today and get each lesson right to your invoice.

Have a great day.

Keep on learning.

Pau Labarta Bajo