Artificial Intelligence has already championed games like Go, chess, quiz-shows, etc. Although AI has outperformed the geniuses, it has yet to show complete comprehension when it comes to understanding the more “practical” side of such games.
The intelligence in AI is something which can be ascertained — right now, it is very different and evolved than the intelligence which was used a decade ago. With this evolving technology, for the first time in history, the Leverhulme Center for the Future of Intelligence in Cambridge and a research institute GoodAI based in Prague have organised a new competition known as the “Animal-AI Olympics”.
The purpose of this competition is to examine and test the autonomous AI agents which are being developed by the top researchers, by placing them into various problem-solving tasks. In this competition, instead of providing a specific task, the AI agents will be provided with a well-defined arena along with a list of cognitive abilities. The competition v1.0 was officially released earlier this month.
Why Animal-AI
Though there is a huge difference between the two subjects, animals and AI, there is one common relation between them — that the researchers in both the fields have been trying to understand the cognition in them for many years. The idea behind the tests is to reveal how it understands, interprets and reasons. Same in the case of AI, researchers are trying to understand from the complex methods how reacts in a certain condition as well as measuring the cognitive abilities which each type of system is most capable of.
Competition Goals
There are certain goals in this competition as mentioned below:
- Benchmark current AI against multiple animal species using a range of established animal cognition tasks
- Present tests towards identifying the cognitive abilities of AI systems
- Determine which AI approaches are most promising for these types of tasks
- Create an ongoing benchmark and data repository for artificial cognition
- Determine which aspects of intelligence are challenging for current AI and which AI already excels at
- Create new experiments to feed back into the animal cognition community that can later be tested with animals
- Bring together two different disciplines to share methods and developments
Final Categories
This competition is far different from the single domain challenges such as Go. It is constituted of 300 different tests and is further divided into 10 final categories and they are food, preferences, obstacles, avoidance, spatial reasoning, generalisation, internal models, object permanence, advanced preferences, and causal reasoning.
Bottom Line
With the release of this competition, the researchers can now start working on the environments and objects. Mathew Crosby, one of the contest’s organisers stated in an interview, “This project captures a lot of elements in AI research that are considered really hard. So far, we haven’t had benchmarks for them, because our benchmarks come from existing games that humans have played in the past. Here, we’re making tasks specifically to test things like generalization and transfer learning. Even if no one does incredibly well in the competition, it will still be useful.”
This competition will indeed help the researchers of artificial intelligence to make a step ahead in this journey. The deadline for final submission of the competition is on 1st November and the results will be declared in the month of December this year. Further, the results will be widely publicised and all the tests will be made available.
Also see:
