Teaching a computer to strafe jump in Quake with reinforcement learning
Channel:
Subscribers:
16,200
Published on ● Video Link: https://www.youtube.com/watch?v=hx7kvTZLHYI
1,504In this video, I use RLLib to train a reinforcement learning agent to strafe jump in the classic video game Quake
Chapters
00:00 Introduction
01:11 What is Reinforcement Learning?
02:42 Quake RL environment
04:41 Results
08:22 Comparison with human WR
09:54 Future direction
Other Videos By Matt's Ramblings
| 2022-09-11 | Minecraft-style logic in Quake |
| 2022-07-04 | Speedrun Science: Beating Quake with code |
| 2022-03-19 | Swapping faces with 100,000 lines of C and 90 lines of Python |
| 2022-01-24 | Quake E1M1 but there are 4096 grunts [4k subscriber celebration] |
| 2021-10-31 | The code behind Quake 3's overbounce bug |
| 2021-09-12 | In search of the perfect speed-drift in Trackmania |
| 2021-06-20 | Quake path traced in Blender |
| 2021-03-13 | Perseverance rover landing footage projected onto map |
| 2021-01-10 | The code behind Quake's movement tricks explained (bunny-hopping, wall-running, and zig-zagging) |
| 2020-12-19 | Better Quake strafe-jumping with genetic algorithms |
| 2020-04-12 | Teaching a computer to strafe jump in Quake with reinforcement learning |
| 2018-03-08 | Quake 3's frame rate dependent physics |
| 2018-02-28 | Strafe-jumping physics explained |
Other Statistics
Quake Statistics For Matt's Ramblings
At this time, Matt's Ramblings has 715,636 views for Quake spread across 13 videos. About an hours worth of Quake videos were uploaded to his channel, roughly 72.94% of the content that Matt's Ramblings has uploaded to YouTube.