Intro AI, Machine Learning Courses Wooing More Students
- By Dian Schaffhauser
Enrollment in artificial intelligence (AI) introductory courses in the United States grew by 3.4 times between 2012 and 2017, and introductory machine learning (ML) classes grew by five times during that same period. That's according to the latest AI Index 2018 Report, a rich collection of data intended to serve as a "comprehensive resource" for anybody interested in the field. The information was contributed by universities, companies, consultancies and associations.
The report observed that ML courses are on a faster trajectory for growth than AI at this point. While the University of California Berkeley's introductory AI course grew by a little under two times between 2012 and 2017, its ML course had 6.8 times as many students. Out of Stanford University's total undergraduate student body, 13 percent enrolled in "Intro to AI" in 2017; an equally high percentage also took the "Intro to ML" course.
But the greatest growth in AI and ML courses didn't take place in this country at all. It occurred in China, where enrollment at Tsinghua University, in particular, grew by 16 times between 2010 and 2017.
The same challenges of gender diversity exist in AI and ML that pervade the rest of computer science. For example, most of those students in the U.S. classes were male. At Stanford, nearly three-quarters of students (74 percent) were male in the AI course and over three-quarters (76 percent) were male in the ML course. At Berkeley, the AI class was 73 percent male and the ML course was 79 percent male.
Among a selection of universities, a comparable dearth of female representation exists in the faculty. The report's authors found that 80 percent of AI professors at representative institutions are male. Among those schools, University College London had the widest gender gap and ETH Zurich had the narrowest gap (though it was still considerable).
The report also examined job openings by AI skill. ML was the most frequently cited job skill in Monster.com job postings overall. However, deep learning grew fastest (35 times) from 2015 to 2017, followed by computer vision, ML techniques, robotics, speech recognition and natural language processing.
Deep learning also drew the largest share of female job applicants among those skills in 2017: 30 percent, according to Gartner TalentNeuron. The gender gap was widest for speech recognition job applicants, where female applicants made up 27 percent of the total. On average, male candidates made up 71 percent of the applicant pool for AI jobs in the United States.
The full report, including information on AI research, investment, "sentiment" of media coverage and myriad other AI activities, is openly available on the AI Index website.
Dian Schaffhauser is a former senior contributing editor for 1105 Media's education publications THE Journal, Campus Technology and Spaces4Learning.