REAL Competition
REAL 2021 Competition– Robot open-Ended Autonomous Learning competition
REAL2021 has started!
Join the competition and submit your robot model code at:
https://eval.ai/web/
Open-ended learning, also named “life-long learning”, “autonomous curriculum learning”, and “no-task learning”, aims to build learning machines and robots that are able to acquire skills and knowledge in an incremental fashion. The REAL competition addresses open-ended learning with a focus on “Robot open-Ended Autonomous Learning” (REAL)”, that is, on systems that: (a) acquire sensorimotor competence that allows them to interact with physical environments, in particular objects; (b) learn in a fully autonomous way, that is, with no human intervention, on the basis of mechanisms such as curiosity, intrinsic motivations, task-free reinforcement learning, self-generated goals, and other mechanisms that might support autonomous learning. The competition has a two-phase structure: during the first “intrinsic phase” your model has a certain time to explore and learn in the environment freely. Then during the second “extrinsic phase” the quality of the knowledge acquired in the intrinsic phase is measured with tasks unknown to the robot during this autonomous phase. You can download a fully open-source software kit from the competition web-site. This allows you to develop increasingly good versions of your model before submitting them directly on our computer and based on a very fast 3D simulator (PyBullet). Moreover, it includes a fully functioning “baseline model” to gear in!
The objectives of REAL are to: – track the state-of-the-art in robot open-ended autonomous learning
– foster research and the proposal of new solutions to the many problems posed by open-ended learning
– develop a benchmark in the field that allows comparing different models based on a specific measure (extrinsic phase performance).
Participation in the competition, which leads to addressing key problems relevant to the ICDL community, is free: everyone is welcome to participate.
The top 3 teams will receive prizes.
For more details on this, see the competition website here.
Important Dates (Updated)
Competition starts: 23rd of August 2021
Competition ends: 14th of June 2022
Final evaluations: 22nd of June 2022
REAL2021 has started!
Join the competition and submit your robot model code at:
https://eval.ai/web/
Open-ended learning, also named “life-long learning”, “autonomous curriculum learning”, and “no-task learning”, aims to build learning machines and robots that are able to acquire skills and knowledge in an incremental fashion. The REAL competition addresses open-ended learning with a focus on “Robot open-Ended Autonomous Learning” (REAL)”, that is, on systems that: (a) acquire sensorimotor competence that allows them to interact with physical environments, in particular objects; (b) learn in a fully autonomous way, that is, with no human intervention, on the basis of mechanisms such as curiosity, intrinsic motivations, task-free reinforcement learning, self-generated goals, and other mechanisms that might support autonomous learning. The competition has a two-phase structure: during the first “intrinsic phase” your model has a certain time to explore and learn in the environment freely. Then during the second “extrinsic phase” the quality of the knowledge acquired in the intrinsic phase is measured with tasks unknown to the robot during this autonomous phase. You can download a fully open-source software kit from the competition web-site. This allows you to develop increasingly good versions of your model before submitting them directly on our computer and based on a very fast 3D simulator (PyBullet). Moreover, it includes a fully functioning “baseline model” to gear in!
The objectives of REAL are to: – track the state-of-the-art in robot open-ended autonomous learning
– foster research and the proposal of new solutions to the many problems posed by open-ended learning
– develop a benchmark in the field that allows comparing different models based on a specific measure (extrinsic phase performance).
Participation in the competition, which leads to addressing key problems relevant to the ICDL community, is free: everyone is welcome to participate.
The top 3 teams will receive prizes.
For more details on this, see the competition website here.
Important Dates (Updated)
Competition starts: 23rd of August 2021
Competition ends: 14th of June 2022
Final evaluations: 22nd of June 2022