We are pushing the boundaries of Artificial Intelligence through the use of principles of generalization. No progress will be made towards a truly general intelligence until we cease creating very narrowly defined AI systems that work on only one niche topic. We aim to produce general Artificial Intelligence Systems that can be applied to a wide array of domains, without the need for human intervention.
We provide a freely available tool to evaluate different AI Systems. This tool runs on opensource tests and wraps them in a simple to use form. The goal is to quickly allow for plug and play between different AI Systems and tests.
We currently have many different flavors of tests used to generate a test suite:
Allen Institute 2
The Obstacle Tower Challenge