Registration deadline: 5th August 2020
- 00 Days
- 00 Hours
- 00 Minutes
- 00 Seconds
Reinforcement Learning Competition (RLCOMP) is a competition relating to Artificial Intelligence which is hosted in Vietnam for the first time. Click here to read the full Regulations
Who Can Join
- Competition is opened to all who can code: Data Scientist, Developer, IT Student,...
- You should know about Information Technology, Artificial Intelligence or Machine Learning
- Experience of Reinforcement Learning is a plus
How To Enter
How to win
- Access to www.codelearn.io to register for an online exam.
- The competition is structured as a Gold Miner game.
Teams compete online with the Organizer's Bots (Screening Stage) or compete directly with each other (Final Stage).
Teams are required to write a program that uses the Reinforcement Learning algorithm to solve the problems set out by the Organizers.
Teams then put their code and trained model (hereinafter referred to as Agent) on the system www.codelearn.io.
The Organizers arrange matches upon reception of Agents from teams.
Each team is required to write a program in Python with popular tools like Tensorflow or Torch based on the Organizers' sample code and use the reinforcement learning algorithm to train the Agent. The Agent, when submitted to the Organizers, will be used to compete on the server system.
By entering the competition you agree to abide by the terms and conditions described below (hereinafter the “Rules”).
Total Value: 500.000.000 VND
- Winner 100.000.000 VND
- 2nd Prize 50.000.000 VND
- 3rd Prize 20.000.000 VND
- Talent Prize 20.000.000 VND
- Teams are allowed to update unlimited Agent versions on the server system, but can only play maximum 20 matches with the Organizer's Bot. (The match with the Bot will be automatically activated after the code is verified successfully).
- For each agent version update, teams will compete in five matches on five given maps.
- In this stage, the positions of the obstacles, boxes of gold and the amount of gold will not change after each match. Each team will compete against three bots of the Organizers (not against other teams).
- The score in the Navigation Challenge stage of a team is the score of the Agent update with the highest amount of gold mined.
- After a maximum of 20 matches with the Bot, the teams with the best score in the Navigation Challenge stage will be qualified to the Treasure Island stage.
- Teams that win the Navigation Challenge stage are randomly divided into groups by the Organizers. Each group has the number of teams by the multiple of four.
- The Organizers sequentially arranges friendly matches for the teams. After each day, teams can view the match history on the system.
- The teams compete online on the system https://codelearn.io
The Final Stage
- Treasure Island consists of 4 rounds: Round 1
- Sudden Death, Round 2
- Moving Forward, Quarter Final
- The True Heir and Final
- Mysterious Treasure.
- The total competition time is 03 hours.
Advisors & Judges
Mr. Nguyen Thanh Nam - Advisor
- Founder of FUNiX
- Former CEO of FPT Corporation
- Former CEO of FPT Software
Mr. Phong Nguyen - Advisor
- AI Researcher
- Visting Researcher at MILA (Montreal Institute for Learning Algorithms) in Canada
- 7 years of AI experience in Hitachi, Tokyo, Japan
Mr. Vu Huu Tiep - Advisor
- Machine Learning Engineer at Google
- Founder of Machine Learning Forum
- Co-Founder of AIviVN
Mr. Hung Ngo - Judge
- Founder of CoTAI (Center of Talent in AI)
- Former AI Director at AINovation
- Former Researcher on AI & Robotics at Stuttgart University
Mr. Nguyen Do Van - Judge
- Director of Research & Development at AI Academy Vietnam
- Adjunct lecturer at University of Engineering and Technology
- Former Researcher at Institute of Military Science and Technology & Reseacrh Center of KDDI, Japan
Mr. Cao Van Viet - Judge
- Founder of Codelearn.io
- Software Engineer Senior at FPT Software