Formula Trinity Autonomous are taking part in the MineRL Diamond Challenge, and they invite you to apply to join them in the Reinforcement Learning group/MineRL team before 21:00 15 July (this Thursday).
You can apply by filling in this form:
Sample efficient learning is an active research challenge that AI agents in Minecraft aim to solve through this competition. Improved RL algorithms would enable robots and AI bots to gain new skills from experience with much less compute than is possible with current state-of-the-art algorithms, essentially democratizing AI for all. Finding diamonds in any Minecraft world is a complex task that is still an interesting challenge for everyone involved!
As part of the team, you will get access to team resources, including generous grants from the MineRL organisers and supporters to support compute costs for our experiments.
You will be involved in the team as part of a long-term goal to bring accessible intelligent robots to Trinity and the wider community. We are a highly-driven student-run team with people from many backgrounds, and we highly support new ideas and suggestions from everyone that is enthusiastic about robotics and AI.
No prior experience in RL is required, though it will help a lot to have an interest in related topics (eg. RL in robotics, sample efficient algorithms, learning from demonstration, multi-task learning).
Challenges and responsibilities:
- Implementing the latest learning algorithms in Python
- Working in both a dynamic project as part of a team, as well as independent research
- Helping the team run experiments and validate results
- Endless curiosity about relevant topics of interest in a fast-moving field
- Communicating and sharing all findings towards informing the wider community about our research
We especially encourage women, and students from all other underrepresented backgrounds, to apply.
We will be responding to applicants shortly after filling out the form, and will do so for all applications sent before the deadline. We will also schedule quick chats with you to discuss your ideas if needed.
Thanks for your time, and looking forward to hearing more from you!
Andrew (Formula Trinity Autonomous)