Persistent understanding of the current state of the field is the minimum and necessary requirement for monitoring of true progress towards general AI
The landscape of AI research is expanding at an accelerating rate. It is increasingly more difficult to keep up-to-date with all relevant progress in the field. We strive to find ways to track progress in the field in a manageable way, for it to be beneficial to all our endeavors and the community at large.
The amount of published research, software libraries, datasets and environments has exploded over the last few years, with hundreds of new papers appearing daily. This project explores ways in which tracking the progress in the field can be made practically more feasible.
A number of researchers, organizations and companies have started to look at this problem, yet existing tools still have significant limitations or narrow scope.
We believe that collaboratively we can discover new ways of mapping the progress of the field in a way that will allow us to digest the increasing amounts of published research, while identifying relevant contributions and avoiding 'noise' in the signal leading towards human-level artificial intelligence.