TOKYO -- In what might be regarded as testament to China's keen interest in artificial intelligence, two of the country's universities were among the leaders in a list showing sources of the most frequently cited research papers in the AI field.
The list, compiled from an analysis of scholarly publications by The Nikkei and Elsevier -- the Amsterdam-based giant of academic publishing -- was created by counting the number of research papers that were cited between 2012 and 2016. A high citation count is generally considered to be an indicator of research quality.
Chinese universities claimed two of the top 10 spots, while U.S. companies like Microsoft and Google also fared well. Japan, which has been slow to embrace AI, lagged far behind China.
Moreover, the top 100 organizations included 15 from China. A major factor behind the country's rapid rise in the AI field is that China has lured back its talented researchers who have studied and worked in the U.S., offering them massive financial incentives to work in a thriving, competitive environment.
The analysis showed that the effects of bolstering the country's research base are feeding into top Chinese universities, as well as many lower-tier universities and institutes. The country's universities are becoming breeding grounds for AI researchers, who are expected to eventually find jobs at large companies and start-ups.
Asian AI aspirants
Singapore also ranked high on the list, with its three institutions finishing in the top 100. Explaining the trend, Yuichiro Anzai, president of the Japan Society for the Promotion of Science, said that the city-state actively recruits talent from overseas.
It has lured researchers from American universities and other organizations with strong AI capabilities, providing them with a good environment for information technology research, including AI.
Malaysia fared relatively well, placing two universities in the top 100. Shigemasa Matsuda, chief researcher at Technology Strategy Center, New Energy and Industrial Technology Development Organization, pointed out that, in addition to their native language, Malaysian researchers speak English, which seems to help them in terms of citation count.
It is comparatively easy for researchers to work in AI if they are able to deal with computers and theory. Because expectations are high that AI could usher in the next industrial revolution, there has been a "rapid improvement in research levels in Asia," according to Matsuda.
Among India's leading institutions, only Jadavpur University cracked the top 200, with the Indian Institute of Technology, Delhi making the top 300. However, most observers agree that the number of research papers in India will increase significantly in the near future, owing to the potential of its IT capabilities.
Korea University, from South Korea, barely made it into the top 100, ranking 97th in the list.
In Japan, only University of Tokyo managed to break the top 100, at 64th. Although 19 Japanese institutions were among the top 500 a decade ago, the latest findings show the number declining to six. This suggests that the country may be unable to catch up in AI research and development, as it lags behind other countries in terms of R&D investment and attracting international research talent.
The dearth of Japanese companies also indicates a lack of technological innovation in the field. Only NTT Group, riding on its strong Japanese-language voice recognition technology, was in the top 500, managing to place 269th.
In a similar list published a decade ago, seven Japanese organizations were in the top 500. Among them was Honda Motor, with its humanoid robot Asimo, finishing at 88th. Sony ranked 201st then, coming on the heels of its marketing Aibo, a robotic dog.