|
Machine Learning |
||
|
AI/Deep Learning Trends
In this page, I am going to gather various statistics about AI/Machine Learning related area from a few well known sources. I would introduce several charts that show big trends (a kind of Mega trends in this area). There would be a couple of reason why you would need to understand. Personally for me, I am interested in tracking these trends mostly for intellectual reason. For any area that I am studying, I like to trace the trends of the technical trends and in many case the understanding on those trends along the time line helps me with the studying the technology itself. In addition, understanding these trends would be helpful if you want to join in this area as your future career. I saw many case where people determine their future career just based on the present popularity (or present demand). But I also saw many cases where they are suffering finding career a few years later when they are elligible to find job espeically for those fast changing area. I think AI/Machine Learning is one of those fast changing (fast evolving area) as of the time I am writing this (Dec 2019). There has been huge improvement for the past several years in terms of technology itself and in terms of the demand in job market, but we have to think of how long further the same trend will maintain in the following years. If you are interested in this area mainly as your future career I would suggest you to think first of how long it will take before you are ready to be in the job market and think what would be the status of this area at the time when you need the job. I agree that AI/Machine Learning is still a lot to be improved and the industry would need new engineers, but I am not sure how long this kind of exponential demand will be maintained. If you are interested in this area, purely as intellectual curiosity.. you don't have to worry about whether it is too early or too late to jump in. Whenever you start, it is right time.
This is what I am tracking in Good Trend. I started tracking a list of keywords as follows : Machine Learning, Deep Learning, Artificial Intelligence, Big Data and Data Science. I am trying to post the update plot once per year as of Dec 1, but you can get the latest trends at any time you want.
In terms of Web Search, the keyword Deep Learning, Big Data, Artificial Intelligence start leveling off this year (or around mid last year) and the keyword Machine Learning and Data Science still maintain the increasing momentum. In terms of YouTube Search, we see drastic downfall for every keyword around the end of 2016, but I think there might be some algorithm changes at Google YouTube statistics at the point.
The statistics in this section is from AI index. From this report, you would have a lot of other aspects of AI trend and I recommend you to go through the original AI index report. It seems that AI index report is published in Dec every year, but they are sharing many of raw datas as they collect.
Refer to the original report here.
NUMBER of AI JOURNAL PUBLICATIONS, 2000-2020
NUMBER of AI PATENT PUBLICATIONS, 2000-2020
AI Index 2019 - Number of AI papers on arXiv (2010-2019)
Comparing the last year, it seems that "CV and Pattern Recognition" and "Artificial Intelligence" papers seems to be level off and "Machine Learning" and "Computation and Language" Papers maintain the exponential growth. (Refer to the full document here)
AI Index 2018 - Number of AI papers on arXiv (2010-2017)
As you see, the number of AI related papers has been being published almost exponentially over the periods of investigation. You would notice that Artificial Intelligence papers seems to go downhill since 2013. But I think this downhill may be because a lot of publishers try to use other hot keywords rather than the term Artificial Intellegence. (Refer to the full document here)
NOTE : It seems very quiet in terms of AI this year, but we got chatGPT and all different types of AIs getting open to public this year.
This year (comparing to last year), a few more keywords came in : Embedded AI, Responsible AI, AI Augmented Developments, Composite AI, Generative AI, AI-Assisted Design.
In this year (comparing to last year), even the term Deep Learning has gone, the area of AI PaaS and Edge AI are still climbing the uphill, and we see new area called Explainable AI, Emotional AI, Adaptive ML.
In this year (comparing to last year), Deep Neural Network(Deep Learing) is still at the peak but the term Machine Learning is gone. Probably the term Machine Learning has become too generic term this year and replaced (or absorbed) by more specific area like AI Paas, Edge AI, Deep Neural Network etc.
In this year, we see two big area of AI, Deep Learning and Machine Learning is around the peak of the curve.
Reference
[2] History, Waves and Winters in AI
|
||
