Hi everyone! I’m excited to share a project I’ve been working on: an AI that learns to balance a ball using reinforcement learning! This is powered by a custom plugin I’ve developed for Unreal Engine, which ports Unity ML-Agents tools to Unreal users.
The plugin aims to make reinforcement learning more accessible to Unreal developers, and I’ll be including the Sphere project as an example to get you started quickly.
The plugin will be available soon, and I’d love to hear your thoughts, feedback, or any ideas for future applications! Here is my discord server if you want to ask me question, learn more about it or get a notification when the plugin is out! : AlanLaboratory
I know that Unreal has it’s own system now, but I think it’s not as well documented or as much as developed as Unity ML Agents. That is why I ported it to Unreal! I think it’s very well documented and easy to use.
The video is a showcase of the plugin! Do not hesitate to share it, comment and subscribe! It would really help me!
Welcome back to the Unreal Engine forums. Hope you’ve been well
This was so genuinely awesome to watch. I’ve never really looked into AI to deeply but this was just so darn interesting. Literally watching Ai “learn” - utterly fascinating. Do you currently have plans to do more tests like this? I know I would personally love to see more
Thank you very much for your kind words! It means a lot for me!
I do have plans to have more videos like this and to share them with the community alongside my plugin. I also have plan to add a story line with my AI Robot called Alan. He will do a lot more than the video I just showed and I will soon publish a video about it. It will not be a demo part of the plugin but will still show what Deep Reinforcement Learning can do.
Do not hesitate to subscribe to the channel if you want to see more, or join the discord if you have questions. I would be more that happy to answer them. The community is small for now, but I do plan to extend it as much as possible.
Thank you so much for your kind words and encouragement! It really means a lot coming from someone on the Unreal team.
I started working on the port of Unity ML-Agents a few months ago because I thought it was such a great tool, and I wanted to bring something similar to Unreal Engine. I genuinely believe Unreal is an amazing platform, and I thought making the transition easier for developers who are already familiar with Unity’s tools would encourage more people to explore Unreal Engine for Deep Reinforcement Learning.
Your Learning Agents system is impressive, and I want to say that I think you and the team are doing an incredible job. I did go through the Unreal tutorial, and I thought it was great—very intuitive and well-executed. At one point, I even considered switching to your plugin, especially since maintaining my port is a lot of work. But by then, I was already quite far into the project, and I felt there was value in continuing to offer this alternative for users who might appreciate the familiarity of the Unity API.
That said, I would absolutely love to support your efforts in any way I can. If you’d like feedback on documentation, tutorials, or anything else, please don’t hesitate to reach out. I’m happy to share my experiences and insights. I do this next to my full-time job, but I’d still be more than willing to help however I can.
Thanks again for taking the time to comment and for all the incredible work you and your team do!
