Reinforcement learning is the class of machine learning algorithms that can be used to solve decision-making problems. It has received a large amount of popular coverage in recent years, because of famous successes such as Google’s AlphaGo beating the human Go champion, and OpenAI’s program competing against teams of humans in DoTA 2. However, this technology has not yet found much traction with real-world industrial problems, apart from a few exceptions such as robotics and autonomous driving.
This talk will give a quick overview of reinforcement learning and its internal working, sticking to mathematical intuition rather than rigorous equations and derivations. Following this, we will highlight what challenges are faced when moving RL from computer simulations and games into the real world. We will focus on potential solutions to these problems, with practical examples from the fields of transportation, logistics, and supply chain operations.
Key Takeaways:
- What is the essence of reinforcement learning, and how to separate its true potential from the media frenzy
- How to identify problems where RL can be truly useful, as opposed to ones where the term sounds cool
- What the technological challenges are when deploying RL in the real world, and how to tackle them
Check out the video below to know more about the talk.
