I've always enjoyed the possibility of reinforcement learning and the utilization of smart agents to execute different tasks. Despite liking it, I was never able to focus on such agents and the whole field of reinforcement learning. Hence, I've decided to start studying them by myself now. I plan to learn a bit every single day. I will start with the theory behind agents, more specifically, problem-solving.
To guide me on this journey, I will use the book Artificial Intelligence: A modern approach by Stuart Russel and Peter Norvig. My initial plan is to follow the steps:
- Study the theory on the book, creating summaries which I'll include in this repository;
- Review the exercises on the text and execute the ones most appropriate to my study;
- Start a toy problem to apply what I've learned.
- CARLA: since my Udacity times, I've been dreaming of using the CARLA simulator to apply RL concepts. At the end of this journey, I want to be able to create even the most simple agent to drive in the CARLA simulator;
- Auto-playing games: I've always loved projects where an agent is used to play a specific game, being Super Mario, Flappy Bird, or even Starcraft.
- Exercises answers: https://aimacode.github.io/aima-exercises/