[DEPRECATED] Visual Maze Solving with Deep Reinforcement Learning in Keras | Detailed Explanation

Опубликовано: 19 Ноябрь 2024
на канале: Jack of Some

10,334

201

2021 Edit: Keras these days no longer has the limitation I talk about here. Take this video with a grain of salt.

This video is about how I built a deep reinforcement learning based visual maze solving networkusing Keras. This is a followup to my second live stream (linked below) where I tried doing this and ran into some issues. I explain the whole system is detail and specifically go over a lot of the mistakes I made that others are likely to make as well.

Code Repo: https://github.com/safijari/jack-of-s...
Slides from David Silver's Excellent Tutorial: https://icml.cc/2016/tutorials/deep_r...
Paper referenced: https://arxiv.org/pdf/1509.06461.pdf

SOME OF MY OTHER VIDOES:
○ Deep RL Stream:    • How to Solve a Basic Reinforcement Le...
○ Vanilla Q Learning Stream:    • How to Solve a Basic Reinforcement Le...
○ Explaining RL to a baby:    • Baby Learns about Artificial Intellig...
○ 5 Common Python Mistakes:    • 5 Things You're Doing Wrong When Prog...
○ Making Python fast:    • Can VSCode be a reasonable Spacemacs ...

Twitter: twitter.com/safijari
Patreon:   / jackofsome