Frequently Asked Questions

I've never heard of XFC, what is it?

XFC is an annual competition in which you build an AI agent to play a game and compete against other teams.

Who can compete?

XFC is open to all students, undergraduate or graduate.

Where can I ask questions not answered here?

We've tried to capture as much info as we can on the website. If there's an issue you can't resolve please feel free to reach out to the organizers. We're all happy to help in anyway we can.

How do I register for XFC?

There's a button on the homepage that will take you to a registration form.

What if I have an issue registering?

If you run into any unforseen issue, please contact the organizers.

What do I need to know to compete?

The language for the competition is Python. Familiarize yourself with this language as well version control with Git. There will be seminars throughout the year to get you up to speed on Fuzzy Logic, Explainable AI, and various machine learing techniques.

Where can I see the agents from previous years?

Find the Previous Agents Here

What OS should I use?

You can program in whatever OS you'd like, the competition will be coded in Python. Know that the organizers will be using Windows and may not be able to accommodate

What resources can I use?

You are free to use whatever resources you can find. Make sure and cite published works or articles you referenced. The only concrete rule for this is you cannot copy code from competitors.

Is there a qualifying round?

Yes. Qualification will involve verifying that your agent can be integrated and run with the competition code without errors. Performance is only considered with respect to the final competition playoff seeding, but will not affect qualification.

What scenarios will be presented?

The AI agents will be presented with a large portfolio of scenarios with varying levels of difficulty and different initial conditions. Some guidance will be given on the specifics and others will be new to the competitors. The goal is to create robust agents that can perform well on scenarios both seen and not seen during training.

How will explainability be evaluated?

• Judges will assess the quality of explanations from the AI.
• Scores will be assigned based on multiple criteria including but not limited to:
  • Ease of understanding
  • Quality of Explanations
  • Sensibility of actions from explanation
  • Innovation

How should I submit my code?

We will fork your code from your team's Github repository. Code must be uploaded to your repo by 11:59pm (EST) on the dates specified in the competition schedule.

How will the agents be scored?

• How many asteroids were hit during the evaluation scenarios
• There may be slight adjustments made to account for specific scenarios

How will the teams be scored?

• Competition Performance (65%)
• Explainability (25%)
• Final Presentation (10%)

How will the competition communicate with the competitors?

We primarily post updates on the ucfuzzycompetition.slack.com channel.

What if I have a question?

We tried to capture as much relevant information as we could on the rules and FAQ pages. if you need further assistance please don't hesitate to reach out to the organizers via email or slack.