AI and Machine Learning research - Can you still manage to keep up?
Updated: Apr 23
The number of #AI and #ML papers on arXiv show no signs of slowing down. @zetavector predicts around 47 thousand new papers this year. Already 2,308 submissions in January 2020. Can you keep up with 128 new titles and abstracts per day?
Every expert working in academia or industry on AI and Machine Learning knows that it is hard to keep track of all the new research that is published. Our numbers show how hard it already is to keep up with only new arXiv submissions. And we're not even counting conferences, journals, blog posts and patents yet...
On arXiv, the computer science rate of submissions started to ramp up in 2007, accelerated to exponential growth, and continued on an even faster growth rate since 2015 (driven largely by the Computer Vision, Machine Learning, and NLP communities). The total number of submissions last year for computer science was 43,317, of which 76.4% were from one of these three domains.
The graph above shows the number of unique submissions per year since 2010*. The yearly average growth rate from 2010 onwards is 43%. If we take the average growth as a predictor for submissions for this year, we should expect more than 47,000 submissions in 2020, or a total of 128 papers per day to keep an eye on. If you only work five days a week, as a mere mortal, that is a stunning average of 180 per day. Fortunately tools like arXiv Sanity or Papers With Code aim to help, but they do not yet dig deeply into the details of the content of each paper.
At Zeta Alpha, we believe that researchers, analysts and decision makers in AI and Machine Learning should harness the power of their own state-of-the-art algorithms to make sense of the research literature. We are building a Deep Learning based NLP, Analytics and Search platform to help keep track of this accelerating growth of the field. Our systems are learning to read and understand, in order to navigate the knowledge around your research project, and to help you make better decisions in your work. We are aiming to release an alpha version in summer 2020, hopefully in time to help you with those 47 thousand arXiv papers for this year!
*) We counted Machine Learning, Artificial Intelligence, Computer Vision and Pattern Recognition, Information Retrieval, Neural and Evolutionary Computing, and Computation and Language categories from arXiv. Keep in mind that when submitting a paper to arXiv the author can select multiple domains, here we counted each paper only once.