The Animal-AI Olympics Testbed
The Animal-AI Olympics Testbed
Embedded Files
Experiments from Animal Cognition
Experiments from Animal Cognition
Click here to play with Version 2.0 of the Animal-AI Environment and watch the agents that participated in the 2019 Animal-AI Olympics competition.
The following experiments are all based on tasks that have been performed on at least one type of animal:
The following experiments are all based on tasks that have been performed on at least one type of animal:
Y-Mazes (matches 69 tests in the testbed)
A Maze in the shape of a Y. One branch contains a preferred reward to the other and usually both can be seen at the same time.
2019 Competition Results
The best score was 61 out of 69. Achieved by cso
The average of the top 10 teams in the competition was 76%
Example Animal Papers:
Pajor, E. A., J. Rushen, and A. M. B. De Passillé. "Dairy cattle’s choice of handling treatments in a Y-maze." Applied Animal Behaviour Science 80.2 (2003)
Pollard, J. C., R. P. Littlejohn, and J. M. Suttie. "Responses of red deer to restraint in a y-maze preference test." Applied Animal Behaviour Science 39.1 (1994)
Tests: [ 2-1-1, 2-1-2, 2-1-3, 2-2-1, 2-2-2, 2-2-3, 2-3-1, 2-3-2, 2-3-3, 2-4-1, 2-4-2, 2-4-3, 2-5-1, 2-5-2, 2-5-3, 2-6-1, 2-6-2, 2-7-1, 2-7-2, 2-7-3, 2-8-1, 2-8-2, 2-8-3, 2-9-1, 2-9-2, 2-9-3, 2-10-1, 2-10-2, 4-7-2, 4-7-3, 4-8-1, 4-8-2, 4-8-3, 4-9-1, 4-9-2, 2-6-3, 4-7-1, 2-12-1, 2-18-1, 2-16-1, 2-10-3, 2-11-1, 2-11-2, 2-11-3, 2-12-2, 2-12-3, 2-13-1, 2-13-2, 2-13-3, 2-14-1, 2-14-2, 2-14-3, 2-15-1, 2-15-2, 2-15-3, 2-16-2, 2-16-3, 2-17-1, 2-17-2, 2-17-3, 2-18-2, 2-18-3, 2-19-1, 2-19-2, 2-19-3, 2-20-1, 2-20-2, 2-20-3, 4-9-3, ]
---
Delayed Gratification (matches 30 tests in the testbed)
The ability to forgo an immediate, less preferred reward for a future, more preferred reward. Solving this robustly in the Animal-AI environment requires understanding that there will be a larger delayed reward based on the physics of the environment.
2019 Competition Results
The best score was 16 out of 30. Achieved by Trrrrr
The average of the top 10 teams in the competition was 34%
Example Animal Paper:
Beran, Michael J. "Maintenance of self-imposed delay of gratification by four chimpanzees (Pan troglodytes) and an orangutan (Pongo pygmaeus)." The Journal of General Psychology 129.1 (2002)
Tests: [ 2-23-3, 2-26-3, 2-26-2, 2-21-1, 2-21-2, 2-21-3, 2-22-1, 2-22-2, 2-22-3, 2-23-1, 2-23-2, 2-24-1, 2-24-2, 2-24-3, 2-25-1, 2-25-2, 2-25-3, 2-26-1, 2-27-1, 2-27-2, 2-27-3, 2-28-1, 2-28-2, 2-28-3, 1-29-1, 2-29-2, 2-29-3, 2-30-1, 2-30-2, 2-30-3, ]
---
Detour Tasks (matches 36 tests in the testbed)
Testing the ability to make a detour around an object to get food and assess the shortest path to the object.
2019 Competition Results
The best score was 12 out of 36. Achieved by Trrrrr
The average of the top 10 teams in the competition was 12%
Example Animal Paper:
Smith, Bradley Philip, and Carla Anita Litchfield. "How well do dingoes, Canis dingo, perform on the detour task?." Animal Behaviour 80.1 (2010)
Tests: [ 3-9-2, 3-7-2, 3-4-1, 3-4-2, 3-4-3, 3-5-1, 3-5-2, 3-5-3, 3-6-1, 3-6-2, 3-6-3, 3-7-1, 3-7-3, 3-8-1, 3-8-2, 3-8-3, 3-9-1, 3-9-3, 3-10-1, 3-10-2, 3-10-3, 3-11-1, 3-11-2, 3-11-3, 3-12-1, 3-12-2, 3-12-3, 4-13-1, 4-13-2, 4-13-3, 4-14-1, 4-14-2, 4-14-3, 4-15-1, 4-15-2, 4-15-3, ]
---
Cylinder Tasks (matches 9 tests in the testbed)
In the testbed this includes both opaque and transparent cylinders
2019 Competition Results
The best score was 7 out of 9. Achieved by ironbar
The average of the top 10 teams in the competition was 48%
Example Animal Paper:
MacLean, Evan L., et al. "The evolution of self-control." Proceedings of the National Academy of Sciences 111.20 (2014)
Tests: [ 3-15-1, 3-13-1, 3-13-2, 3-13-3, 3-14-1, 3-14-2, 3-14-3, 3-15-2, 3-15-3, ]
---
Thorndike escape experiments (matches 18 tests in the testbed)
These are recreations (of differing realism) of Thordike's experiments on cats, dogs, and chicks where the agent must escape from a confined area and food is placed outside.
2019 Competition Results
The best score was 9 out of 18. Achieved by sirius, DeepFox
The average of the top 10 teams in the competition was 27%
Example Animal Paper:
Thorndike, Edward. Animal intelligence Experimental studies. Routledge, (2017)
Tests: [ 3-27-3, 3-25-1, 3-25-2, 3-25-3, 3-26-1, 3-26-2, 3-26-3, 3-27-1, 3-27-2, 3-28-1, 3-28-2, 3-28-3, 3-29-1, 3-29-2, 3-29-3, 3-30-1, 3-30-2, 3-30-3, ]
---
T-Maze (matches 9 tests in the testbed)
Like a Y-Maze except that both arms are not visible at the same time.
2019 Competition Results
The best score was 9 out of 9. Achieved by Juramaia, cso, Trrrrr
This category was SOLVED!
The average of the top 10 teams in the competition was 66%
Example Animal Paper:
Qin, Jianhua, and Aaron R. Wheeler. "Maze exploration and learning in C. elegans." Lab on a Chip 7.2 (2007)
Tests: [ 5-1-2, 5-1-3, 5-2-1, 5-2-3, 5-3-1, 5-3-2, 5-1-1, 5-3-3, 5-2-2, ]
---
Spatial Elimination (matches 27 tests in the testbed)
Spatial properties can be used to infer the location of food. E.g. it can't be in the open space so if there is any it must be behind that wall.
2019 Competition Results
The best score was 7 out of 27. Achieved by DeepFox
The average of the top 10 teams in the competition was 17%
Example Animal Paper:
Herrmann, Esther, et al. "Humans have evolved specialized skills of social cognition The cultural intelligence hypothesis." science 317.5843 (2007)
Tests: [ 5-4-1, 5-4-2, 5-4-3, 5-5-1, 5-5-2, 5-6-1, 5-6-2, 5-6-3, 5-7-1, 5-7-2, 5-7-3, 5-8-1, 5-8-2, 5-8-3, 5-9-1, 5-9-2, 5-9-3, 5-10-1, 5-10-2, 5-10-3, 5-11-1, 5-11-3, 5-12-1, 5-12-2, 5-12-3, 5-5-3, 5-11-2, ]
---
Support and Gravity Bias (matches 27 tests in the testbed)
Tasks involving gravity and food supported on other objects.
2019 Competition Results
The best score was 12 out of 27. Achieved by ironbar, BronzeBlood
The average of the top 10 teams in the competition was 31%
Example Animal Paper:
Hauser, Marc D., et al. "What guides a search for food that has disappeared? Experiments on cotton-top tamarins (Saguinus oedipus)." Journal of Comparative Psychology 115.2 (2001)
Tests: [ 5-13-1, 5-13-2, 5-14-1, 5-14-2, 5-14-3, 5-15-1, 5-15-2, 5-15-3, 5-16-1, 5-16-2, 5-16-3, 5-17-1, 5-17-2, 5-17-3, 5-18-1, 5-18-2, 5-18-3, 5-19-2, 5-19-3, 5-20-1, 5-20-2, 5-20-3, 5-21-1, 5-21-2, 5-21-3, 5-13-3, 5-19-1, ]
---
Radial Mazes (matches 27 tests in the testbed)
Mazes with a number of spokes radiating out from a central hub.
2019 Competition Results
The best score was 16 out of 27. Achieved by Trrrrr
The average of the top 10 teams in the competition was 29%
Example Animal Paper:
Hughes, Roger N., and Christine M. Blight. "Algorithmic behaviour and spatial memory are used by two intertidal fish species to solve the radial maze." Animal Behaviour 58.3 (1999)
Tests: [ 5-22-1, 5-22-2, 5-23-1, 5-23-2, 5-23-3, 5-24-1, 5-24-2, 5-24-3, 5-25-2, 5-25-3, 5-26-1, 5-26-2, 5-26-3, 5-27-1, 5-27-2, 5-27-3, 5-28-1, 5-28-2, 5-28-3, 5-29-1, 5-29-2, 5-29-3, 5-30-1, 5-30-2, 5-22-3, 5-25-1, 5-30-3, ]
---
Object Permanence (matches 90 tests in the testbed)
These tests all involve food that moves out of sight that the agent needs to still attain.
2019 Competition Results
The best score was 23 out of 90. Achieved by ironbar
The average of the top 10 teams in the competition was 9%
Example Animal Papers:
Chiandetti, Cinzia, and Giorgio Vallortigara. "Intuitive physical reasoning about occluded objects by inexperienced chicks." Proceedings of the Royal Society B Biological Sciences 278.1718 (2011)
Herrmann, Esther, et al. "Humans have evolved specialized skills of social cognition The cultural intelligence hypothesis." science 317.5843 (2007)
Tests: [ 8-19-2, 8-19-3, 8-20-1, 8-20-3, 8-21-1, 8-21-2, 8-21-3, 8-22-1, 8-22-2, 8-22-3, 8-23-1, 8-23-2, 8-23-3, 8-24-1, 8-24-3, 8-25-1, 8-25-2, 8-25-3, 8-26-1, 8-26-3, 8-27-1, 8-27-2, 8-27-3, 8-28-1, 8-28-2, 8-28-3, 8-29-1, 8-29-2, 8-29-3, 8-30-1, 8-30-2, 8-30-3, 8-7-3, 8-8-1, 8-8-2, 8-8-3, 8-9-1, 8-9-2, 8-9-3, 8-10-1, 8-10-2, 8-10-3, 8-11-1, 8-11-3, 8-12-1, 8-12-2, 8-12-3, 8-13-1, 8-13-2, 8-13-3, 8-14-1, 8-14-2, 8-14-3, 8-15-1, 8-15-2, 8-15-3, 8-16-1, 8-16-2, 8-16-3, 8-17-2, 8-17-3, 8-18-1, 8-18-2, 8-18-3, 8-19-1, 8-24-2, 8-20-2, 8-26-2, 8-11-2, 8-17-1, 8-1-1, 8-1-2, 8-1-3, 8-2-1, 8-2-2, 8-2-3, 8-3-1, 8-3-2, 8-3-3, 8-4-1, 8-4-2, 8-4-3, 8-5-1, 8-5-2, 8-5-3, 8-6-1, 8-6-2, 8-6-3, 8-7-1, 8-7-2, ]
---
Numerosity (matches 90 tests in the testbed)
These tests all involve counting to navigate to the compartment with the most food.
2019 Competition Results
The best score was 45 out of 90. Achieved by BronzeBlood, Oltau.ai, sungbinchoi
The average of the top 10 teams in the competition was 43%
Example Animal Paper:
Herrmann, Esther, et al. "Humans have evolved specialized skills of social cognition The cultural intelligence hypothesis." science 317.5843 (2007)
Tests: [ 9-5-1, 9-5-2, 9-5-3, 9-6-2, 9-6-3, 9-7-1, 9-7-2, 9-7-3, 9-8-1, 9-8-2, 9-6-1, 9-8-3, 9-18-3, 9-29-1, 9-9-1, 9-14-3, 9-20-3, 9-26-2, 9-29-2, 9-1-1, 9-1-2, 9-1-3, 9-2-1, 9-2-2, 9-2-3, 9-3-1, 9-3-2, 9-3-3, 9-4-1, 9-4-2, 9-4-3, 9-9-2, 9-9-3, 9-10-1, 9-10-2, 9-10-3, 9-11-1, 9-11-2, 9-11-3, 9-12-1, 9-12-2, 9-12-3, 9-13-1, 9-13-2, 9-13-3, 9-14-1, 9-14-2, 9-15-1, 9-15-2, 9-15-3, 9-16-1, 9-16-2, 9-16-3, 9-17-1, 9-17-2, 9-17-3, 9-18-1, 9-18-2, 9-19-1, 9-19-2, 9-19-3, 9-20-1, 9-20-2, 9-21-1, 9-21-2, 9-21-3, 9-22-1, 9-22-2, 9-22-3, 9-23-1, 9-23-2, 9-23-3, 9-24-1, 9-24-2, 9-24-3, 9-25-1, 9-25-2, 9-25-3, 9-26-1, 9-26-3, 9-27-1, 9-27-2, 9-27-3, 9-28-1, 9-28-2, 9-28-3, 9-29-3, 9-30-1, 9-30-2, 9-30-3, ]
---
Tool Use (matches 90 tests in the testbed)
These test are based on the ability to use the pushable objects in the arena as makeshift tools to get food. They are the most complicated in the testbed and extend to the ability to perform simple causal reasoning about the outcome of actions.
2019 Competition Results
The best score was 10 out of 90. Achieved by DeepFox
The average of the top 10 teams in the competition was 2%
Example Animal Papers:
Jacobs, Ivo F., and Mathias Osvath. "The string-pulling paradigm in comparative psychology." Journal of Comparative Psychology 129.2 (2015)
Hauser, Marc David, Jerald Kralik, and Carezza Botto-Mahan. "Problem solving and functional design features experiments on cotton-top tamarins, Saguinus oedipus oedipus." Animal Behaviour 57.3 (1999)
Epstein, R., et al. Insight in the pigeon antecedents and determinants of an intelligent performance." Nature 308.5959 (1984)
Mulcahy, Nicholas J., and Josep Call. "How great apes perform on a modified trap-tube task." Animal cognition 9.3 (2006)
Tests: [ 10-14-3, 10-10-3, 10-16-3, 10-22-2, 10-1-1, 10-1-2, 10-1-3, 10-2-1, 10-2-2, 10-2-3, 10-3-1, 10-3-2, 10-3-3, 10-4-1, 10-4-2, 10-4-3, 10-5-1, 10-5-2, 10-5-3, 10-6-1, 10-6-2, 10-6-3, 10-7-1, 10-7-2, 10-7-3, 10-8-1, 10-8-2, 10-8-3, 10-9-1, 10-9-2, 10-9-3, 10-10-1, 10-10-2, 10-11-1, 10-11-2, 10-11-3, 10-12-1, 10-12-2, 10-12-3, 10-13-1, 10-13-2, 10-13-3, 10-14-1, 10-14-2, 10-15-1, 10-15-2, 10-15-3, 10-16-1, 10-16-2, 10-17-1, 10-17-2, 10-17-3, 10-18-1, 10-18-2, 10-18-3, 10-19-1, 10-19-2, 10-19-3, 10-20-1, 10-20-2, 10-20-3, 10-21-1, 10-21-2, 10-21-3, 10-22-1, 10-22-3, 10-23-1, 10-23-2, 10-23-3, 10-24-1, 10-24-2, 10-24-3, 10-25-1, 10-25-2, 10-25-3, 10-26-1, 10-26-2, 10-26-3, 10-27-1, 10-27-2, 10-27-3, 10-28-1, 10-28-2, 10-28-3, 10-29-1, 10-29-2, 10-29-3, 10-30-1, 10-30-2, 10-30-3, ]
---
Introductory experiments
Introductory experiments
These are simple experiments designed to introduce a feature of the environment that the agent must habituate to
Basic Food Ahead (matches 9 tests in the testbed)
Very simple tasks with basic food directly in front of the agent.
2019 Competition Results
The best score was 9 out of 9. Achieved by BronzeBlood, Oltau.ai, sungbinchoi, DeepFox, ARF-RL, UniboTeam, Juramaia, cso, mmIA, Gyutan, winter2109, forest, CUMIN, Qodiak, y.yang, doot, KMU-AIL, BLAI, CHROMA, jinrohs, TheAnimalsEscapedFromPenn, AirbrainNM, Koozyt_AnimalAI_alpha, Juohmaru, Thursday, hhq126152, Trrrrr, ironbar, sirius
This category was SOLVED!
The average of the top 10 teams in the competition was 100%
Tests: [ 1-1-1, 1-1-2, 1-1-3, 1-2-1, 1-2-2, 1-2-3, 1-3-1, 1-3-2, 1-3-3, ]
---
Basic Navigate to Food (matches 6 tests in the testbed)
Very simple tasks to navigate to visible food.
2019 Competition Results
The best score was 6 out of 6. Achieved by BronzeBlood, Oltau.ai, sungbinchoi, DeepFox, ARF-RL, UniboTeam, Juramaia, cso, mmIA, Gyutan, winter2109, forest, CUMIN, y.yang, animalAI-challenge, GoGoAI, doot, KMU-AIL, BLAI, CHROMA, jinrohs, TheAnimalsEscapedFromPenn, AirbrainNM, Koozyt_AnimalAI_alpha, Juohmaru, Thursday, Trrrrr, ironbar, sirius
This category was SOLVED!
The average of the top 10 teams in the competition was 100%
Tests: [ 1-4-1, 1-4-2, 1-4-3, 1-5-1, 1-5-2, 1-5-3, ]
---
Basic Food Variations (matches 18 tests in the testbed)
Navigating to visible food of different types.
2019 Competition Results
The best score was 18 out of 18. Achieved by BronzeBlood, Oltau.ai, sungbinchoi, DeepFox, ARF-RL, UniboTeam, cso, mmIA, Gyutan, animalAI-challenge, GoGoAI, CHROMA, Koozyt_AnimalAI_alpha, daydayup, Juohmaru, Trrrrr, ironbar, sirius
This category was SOLVED!
The average of the top 10 teams in the competition was 99%
Tests: [ 1-6-2, 1-6-1, 1-6-3, 1-7-1, 1-7-2, 1-7-3, 1-8-1, 1-8-2, 1-8-3, 1-9-1, 1-9-2, 1-9-3, 1-10-1, 1-10-2, 1-10-3, 1-11-1, 1-11-2, 1-11-3, ]
---
Basic Exploration (matches 18 tests in the testbed)
Empty arenas with a single food of varying size and type. Not necessarily initially visible.
2019 Competition Results
The best score was 18 out of 18. Achieved by DeepFox, Gyutan, winter2109, CUMIN, ironbar
This category was SOLVED!
The average of the top 10 teams in the competition was 92%
Tests: [ 1-12-1, 1-16-1, 1-12-2, 1-12-3, 1-13-1, 1-13-2, 1-13-3, 1-14-1, 1-14-2, 1-14-3, 1-15-1, 1-15-2, 1-15-3, 1-16-2, 1-16-3, 1-17-1, 1-17-2, 1-17-3, ]
---
Basic Multiple Food (matches 9 tests in the testbed)
Empty arenas with many yellow foods that have to all be collected
2019 Competition Results
The best score was 9 out of 9. Achieved by Juramaia, Trrrrr, sirius
This category was SOLVED!
The average of the top 10 teams in the competition was 83%
Tests: [ 1-20-1, 1-20-2, 1-20-3, 1-21-1, 1-21-2, 1-21-3, 1-22-1, 1-22-2, 1-22-3, ]
---
Basic Food and Obstacles (matches 9 tests in the testbed)
The introductory tests that may contain obstacles that need to be navigated around to reach the food.
2019 Competition Results
The best score was 8 out of 9. Achieved by UniboTeam, Qodiak, Trrrrr, ironbar
The average of the top 10 teams in the competition was 53%
Tests: [ 3-1-1, 3-1-3, 3-1-2, 3-3-2, 3-2-1, 3-2-2, 3-2-3, 3-3-1, 3-3-3, ]
---
All other experiments in the testbed
All other experiments in the testbed
Moving Food (matches 9 tests in the testbed)
Siple arenas with a single moving piece of food.
2019 Competition Results
The best score was 7 out of 9. Achieved by Oltau.ai, sungbinchoi, KMU-AIL, jinrohs, Yossy, Trrrrr, ironbar
The average of the top 10 teams in the competition was 71%
Tests: [ 1-18-1, 1-18-2, 1-18-3, 1-19-1, 1-19-2, 1-19-3, 4-30-1, 4-30-2, 4-30-3, ]
---
Unreachable Food (matches 6 tests in the testbed)
Problems where there is only negative reward
2019 Competition Results
The best score was 6 out of 6. Achieved by Juramaia, Gyutan, forest, GoGoAI, Koozyt_AnimalAI_alpha, ironbar
This category was SOLVED!
The average of the top 10 teams in the competition was 72%
Tests: [ 4-28-3, 1-24-1, 1-24-2, 1-24-3, 4-28-1, 4-28-2, ]
---
Multiple Food Stationary (matches 9 tests in the testbed)
Empty arenas with all types of food. Yellow needs to be collected before Green.
2019 Competition Results
The best score was 7 out of 9. Achieved by Trrrrr, sirius
The average of the top 10 teams in the competition was 47%
Tests: [ 1-23-3, 1-23-1, 1-23-2, 1-27-1, 1-27-2, 1-27-3, 1-28-1, 1-28-2, 1-28-3, ]
---
Multiple Food Moving (matches 9 tests in the testbed)
Empty arenas with all types of food. Yellow needs to be collected before Green.
2019 Competition Results
The best score was 5 out of 9. Achieved by sirius, winter2109
The average of the top 10 teams in the competition was 28%
Tests: [ 1-26-3, 1-26-2, 1-26-1, 1-29-1, 1-29-2, 1-29-3, 1-30-1, 1-30-2, 1-30-3, ]
---
Avoid Red (matches 36 tests in the testbed)
Have to navigate around negative rewards to get the food.
2019 Competition Results
The best score was 18 out of 36. Achieved by ironbar
The average of the top 10 teams in the competition was 19%
Tests: [ 4-1-1, 4-1-3, 4-2-1, 4-2-2, 4-2-3, 4-3-1, 4-3-2, 4-3-3, 4-1-2, 4-22-3, 1-25-1, 1-25-2, 1-25-3, 4-12-1, 4-12-2, 4-12-3, 4-19-1, 4-19-2, 4-19-3, 4-20-1, 4-20-2, 4-20-3, 4-21-1, 4-21-2, 4-21-3, 4-22-1, 4-22-2, 4-23-1, 4-23-2, 4-23-3, 4-24-1, 4-24-2, 4-24-3, 4-29-1, 4-29-2, 4-29-3, ]
---
Ramp Usage (matches 27 tests in the testbed)
In these tasks the agent must use a ramp to gain access to part of the environment
2019 Competition Results
The best score was 10 out of 27. Achieved by ironbar
The average of the top 10 teams in the competition was 9%
Tests: [ 3-16-1, 3-16-2, 3-18-1, 3-18-2, 3-18-3, 3-19-1, 3-19-2, 3-19-3, 3-20-1, 3-17-3, 4-26-3, 3-16-3, 3-17-1, 3-17-2, 3-20-2, 3-20-3, 3-21-1, 3-21-2, 3-21-3, 4-25-1, 4-25-2, 4-25-3, 4-26-1, 4-26-2, 4-27-1, 4-27-2, 4-27-3, ]
---
Pushing Boxes (matches 9 tests in the testbed)
In these tasks the agent must push boxes in order to gain access to part of the environment
2019 Competition Results
The best score was 4 out of 9. Achieved by CUMIN
The average of the top 10 teams in the competition was 13%
Tests: [ 3-22-1, 3-22-2, 3-22-3, 3-23-1, 3-23-2, 3-23-3, 3-24-1, 3-24-2, 3-24-3, ]
---
Hot Zones (matches 24 tests in the testbed)
Experiments involving the hot zones, where the agent receives increased negative reward per time spent
2019 Competition Results
The best score was 18 out of 24. Achieved by Oltau.ai, sungbinchoi, ARF-RL
The average of the top 10 teams in the competition was 62%
Tests: [ 4-4-1, 4-4-2, 4-4-3, 4-5-1, 4-5-2, 4-5-3, 4-6-1, 4-6-2, 4-6-3, 4-17-1, 4-16-3, 4-10-1, 4-10-2, 4-10-3, 4-11-1, 4-11-2, 4-11-3, 4-16-1, 4-16-2, 4-17-2, 4-17-3, 4-18-1, 4-18-2, 4-18-3, ]
---
Genralisation and Adaptability (matches 90 tests in the testbed)
This category tested the agent's ability to adapt when the colours of the objects in the environment were changed (food was left the same).
2019 Competition Results
The best score was 49 out of 90. Achieved by Trrrrr
The average of the top 10 teams in the competition was 36%
Tests: [ 6-9-2, 6-9-3, 6-10-1, 6-10-2, 6-10-3, 6-11-1, 6-11-2, 6-11-3, 6-12-1, 6-12-2, 6-12-3, 6-13-1, 6-13-2, 6-13-3, 6-14-1, 6-14-2, 6-15-1, 6-15-2, 6-15-3, 6-16-1, 6-16-2, 6-16-3, 6-17-1, 6-17-2, 6-17-3, 6-18-1, 6-18-2, 6-19-1, 6-19-2, 6-19-3, 6-20-1, 6-20-2, 6-21-1, 6-21-2, 6-21-3, 6-22-1, 6-22-2, 6-22-3, 6-23-1, 6-23-2, 6-23-3, 6-24-1, 6-24-2, 6-24-3, 6-25-1, 6-25-2, 6-25-3, 6-26-1, 6-26-3, 6-27-1, 6-27-2, 6-27-3, 6-28-1, 6-28-2, 6-28-3, 6-29-2, 6-29-3, 6-30-1, 6-30-3, 6-8-3, 6-18-3, 6-29-1, 6-30-2, 6-9-1, 6-14-3, 6-20-3, 6-26-2, 6-1-1, 6-1-2, 6-1-3, 6-2-1, 6-2-2, 6-2-3, 6-3-1, 6-3-2, 6-3-3, 6-4-1, 6-4-2, 6-4-3, 6-5-1, 6-5-2, 6-5-3, 6-6-1, 6-6-2, 6-6-3, 6-7-1, 6-7-2, 6-7-3, 6-8-1, 6-8-2, ]
---
Internal Models Introduction (matches 90 tests in the testbed)
These test all involve the lights going out in the environment at some point.
2019 Competition Results
The best score was 52 out of 90. Achieved by Trrrrr
The average of the top 10 teams in the competition was 44%
Tests: [ 7-26-3, 7-27-1, 7-27-2, 7-27-3, 7-28-1, 7-28-2, 7-28-3, 7-29-1, 7-29-2, 7-29-3, 7-30-1, 7-30-2, 7-30-3, 7-4-1, 7-4-2, 7-4-3, 7-5-1, 7-5-2, 7-5-3, 7-6-1, 7-6-3, 7-7-1, 7-7-2, 7-7-3, 7-8-1, 7-8-2, 7-8-3, 7-9-1, 7-9-2, 7-9-3, 7-10-1, 7-10-2, 7-10-3, 7-11-1, 7-11-2, 7-11-3, 7-12-3, 7-13-1, 7-13-2, 7-13-3, 7-14-1, 7-14-2, 7-14-3, 7-15-1, 7-15-2, 7-15-3, 7-16-1, 7-16-2, 7-16-3, 7-17-1, 7-17-2, 7-17-3, 7-18-2, 7-18-3, 7-19-1, 7-19-2, 7-19-3, 7-20-1, 7-12-1, 7-26-2, 7-6-2, 7-12-2, 7-18-1, 7-23-3, 7-1-1, 7-1-2, 7-1-3, 7-2-1, 7-2-2, 7-2-3, 7-3-1, 7-3-2, 7-3-3, 7-20-2, 7-20-3, 7-21-1, 7-21-2, 7-21-3, 7-22-1, 7-22-2, 7-22-3, 7-23-1, 7-23-2, 7-24-1, 7-24-2, 7-24-3, 7-25-1, 7-25-2, 7-25-3, 7-26-1, ]
--- End
Page updated
Report abuse