1. Background of the Event

To enhance Yuntech students' understanding of artificial intelligence and to inspire more students to venture into the realm of AI, this competition is specially organized. Through this event, participants will become acquainted with the cloud computing platforms widely utilized in the industry, gain insights into AI-related technologies, and explore the popular applications of AI in autonomous vehicles.

2. Organizers

1. Organizer: Yuntech Department of Information Management
2. Co-organizer: Yuntech Teaching Excellence Center
3. Contact Person: Professor Arthur Chang, Department of Information Management, Yuntech, Email住址會使用灌水程式保護機制。你需要啟動Javascript才能觀看它

3. Competition Regulations

1. Registration Period: From now until October 21st, 2025.
2. Team Formation: Participants must be currently enrolled students of Yuntech. Each team shall consist of 1 to 4 members. Upon registration, a photo of both sides of the student ID must be uploaded. Once verified, the organizer will send an invitation email to the registered email address, granting access to the AI Console/AI Stack cloud computing platform, enabling the development of their own reinforcement learning models.
3. Scoring Method: Two online racing point competitions will be held. The top ten finishers in each race will be awarded points as follows: 25, 18, 15, 12, 10, 8, 6, 4, 2, and 1 point, respectively.
4. Ranking Calculation: The warm-up race (October 22nd-24th) will not count towards points. Rankings will be based on the total points accumulated from the two-point races (October 28th and 29th). In the event of a tie, the total race time will determine the ranking, with the shorter time prevailing.
5. Competition Restrictions: Only self-developed Reinforcement Learning models may be submitted to the competition platform. Sharing or using models developed by others is strictly prohibited.
   1. Venue: The competition will take place on designated virtual race tracks. Each race’s track map will be disclosed before the event, featuring curves and straight sections to simulate real-world driving challenges.
   2. Format:
      • Each race requires completion of 3 laps to be considered valid, conducted in a timed trial format.
      • Participants may submit models up to a specified limit; the system will retain the model with the best performance as the final result.
      • Race environment files will be provided to participants for model training.
   3. Out-of-Bounds Handling: Each time a vehicle goes out of bounds, a 3-second penalty will be added, and the car will be reset to the starting point to resume driving.
   4. Checkpoint and Lap Counting:
      • Vehicles must sequentially pass all checkpoints to accumulate laps.
      • Failure to pass checkpoints in order will result in a reset to the correct checkpoint position.
   5. Fairness and Anti-Cheating Measures:
      • The use of external tools or the manipulation of the competition is strictly prohibited.
      • Vehicle behavior will be monitored during the race; any anomalies detected will result in the disqualification of the model.
   6. Competition Platform:
      • The platform will save the best results of each attempt for participants to review and compare.
      • Model submissions shall be made on a team basis.
   7. Training Requirements: Please use PyTorch version 2.1.0. The learning method is unrestricted but must comply with the submission format specifications.

4. Awards

1. First Place: Gift vouchers worth NT$5,000 and a certificate of merit;
2. Second Place: Gift vouchers worth NT$3,000 and a certificate of merit;
3. Third Place: Gift vouchers worth NT$2,000 and a certificate of merit.

5. Notes

The organizer reserves the right to amend any regulations about this competition. The Higher Education Sprout Project funds this event.

link: https://airacing.yuntech.edu.tw