AI Olympics Competition

Participation

Background

As reinforcement learning gradually becoming the hottest topic in the academic field of artificial intelligence, the Institute of Automation, Chinese Academy of Sciences will hold an agent competition - AI Olympics Competition, to further promote the development of reinforcement learning research.

The purpose of this competition is to provide a competitive and communication platform for domestic and foreign agent algorithm researchers, and to promote the research and application of decision-making intelligence with reinforcement learning as the technical path. Different from the traditional data-based perceptual intelligence competition, the agent competition based on the interaction between the agent and the environment has a more cutting-edge academic meaning and higher ornamental.

The competition was hosted by the Institute of Automation, Chinese Academy of Sciences, and co-organized by the Institute for Artificial Intelligence, Peking University, APEX Data and Knowledge Management Lab of Shanghai Jiao Tong University, Shanghai Digital Brain Research Institute, Shanghai Boyu Education Technology Co., Ltd. and the RLChina Reinforcement Learning Community.

Committee

John E. Hopcroft

Professor of Cornell University, winner of Turing Award and foreign academician of Chinese Academy of Sciences

Xiaotie Deng

Professor of Peking University, foreign academician of the European Academy of Sciences

Jun Wang

Professor of University College London, Turling Fellow

Wenxin Li

Professor of Peking University, vice president of Institute for Artificial Intelligence, Peking University

Yang Yu

Professor of Nanjing University

Bo An

Associate Professor of Nanyang Technological University

Jianye Hao

Associate Professor of Tianjin University, Noah's Ark Lab, Huawei

Weinan Zhang

Associate Professor of Shanghai Jiao Tong University

Haifeng Zhang

Associate Professor of Institute of Automation, Chinese Academy of Sciences

Environment

We provide the integrated Olympics environment as competition envionrment, which combines different subgames. The subgames are running, table hockey, football, wrestling, etc.

  • In running, the goal is to reach the end as fast as possible in a randomly selected map.
  • In table hockey, agents need to hit the ball into the opponent's goal through collision and prevent their own goal from being broken, but they can only move freely in their own half court.
  • In football, agents share the same objective as in table hockey except that they can move freely on the court.
  • In wrestling, the agent who knocks others out of bounds while keeping staying in bounds wins the game.
  • The evaluation of the competition is going to use this envionrment to run different subgames in random order to test the submitted agents.

Prize

First Prize: 1 team

￥25000/team

Second Prize: 3 teams

￥15000/team

Third Prize: 6 teams

￥5000/team

Participation

Come and Sign up for the competition!

Participation


In order to help students systematically learn reinforcement learning and get better results in this competition, we provide the "Hands-on Reinforcement Learning" course produced by Associate Professor Weinan Zhang of Shanghai Jiao Tong University and Dr. Jian Shen of APEX Laboratory of Shanghai Jiao Tong University, which is as follows:

  • Login in on the PC https://www.qingzhouzhixue.com Qingzhou Zhixue intelligent skill learning platform
  • Click【Start your learning journey】
  • Select【Hands-on Reinforcement Learning】learning path
  • Enter the activation code【jidi2022】
  • Conduct a preschool assessment and start learning based on the learning path generated by the platform

Timeline

The timeline and format of this competition may be changed, and the final interpretation rights belong to the organizer.

Open Registration: 2022.4.20 to 2022.6.16 10:30 (UTC+8)
Warm-up 1st Round (Submission Deadline 2022.5.5 10:30 UTC+8): Swiss Round double competition system, the result is not included in the final competition result
Warm-up 2st Round (Submission Deadline 2022.5.17 10:30 UTC+8): Swiss Round double competition system, the result is not included in the final competition result
Warm-up 3rd Round (Submission Deadline 2022.5.31 10:30 UTC+8): Swiss Round double competition system, the result is not included in the final competition result
Warm-up 4th Round (Submission Deadline 2022.6.7 10:30 UTC+8): Swiss Round double competition system, the result is not included in the final competition result
Final Round (Submission Deadline 2022.6.16 10:30 UTC+8): Swiss Round double competition system, the score will be the final total score

Sponsorship

上海数字大脑科技研究院有限公司

Recruiter: hr@digitalbrain.cn