The China Academic Perspective On Artificial Intelligence & Big Data
As part of my Eisenhower Fellowship to China for the month of June, I will be meeting with government agencies, academic institutions, companies, and regular citizens across 6 cities in China to learn more about the state of technology in China. To start our fellowship, Renmin University and the China Education Association for International Exchange organized a series of lectures on a variety of topics — from policy-making to how China is tackling an aging population. Two sessions that stood out for me this week were tied to the work of two professors that covered research in artificial intelligence and the big data ecosystem in China. Here’s a quick breakdown of some of what we learned during these sessions.
Artificial Intelligence Tackles Video
First up was Qin Jin, a professor at the School of Information, Renmin University of China. Professor Qin explained her work was focused on using artificial intelligence to create a Multi-Aspect, Multi-Level, Multi-modal Analysis and Understanding of Video. Essentially, her team had blended a few different types of artificial intelligence (neural networks, machine learning, etc.) to conduct a contextual analysis of video in what she referred to as Multimodality Fusion. This research was further extended to analyzing the emotional context of videos leveraging much of the same techniques, which is also something that many researchers in the United States have also been working on. This research will further enable consumer-targeted artificial intelligence systems to better understand and respond to users and the context of their interactions.
After a series of questions and answers, here are some additional takeaways based on our discussion on artificial intelligence.
- AI in China is commercially available to developers, similar to how AWS, IBM, Google and other big companies in the United States provide commercial access to AI capabilities as a service and typically through an application programming interface (API). The major AI-as-a-service players in China are Baidu, Tencent, Alibaba.
- China is experiencing the same challenges around having enough data to fully utilize neural networks and is also exploring capsule networks as a potential solution.
- Voice assistants are also growing in popularity in China, and there are similar growing fears around privacy and data utilization.
Big Data In China
Next up was Professor Hanfang Yang, Associate Professor at the School of Statistics at the Renmin University of China. Professor Hanfang, explained the open data ecosystem in China.
The big data infrastructure in China is very similar to that of the United States with a few exceptions. Some of the key big data infrastructure technology components (CPU, GPUs, etc.) in China are currently obtained through international trade, much of which is from the United States. Our technology ecosystems for big data are interconnected and mutually supportive. The amount of data produced has exponentially increased due to the highly-competitive nature of the Chinese cloud market place and the decreasing cost of cellphones (and other IoT devices). In fact, some of our group of fellows purchased new, relatively state-of-the-art, Chinese cell phones for just under 1,000 RMB — which is approximately $144 US-dollars as I write this. There are also numerous infrastructure improvements underway, including the rollout of 5G wireless technology. Just like in the United States, 5G is in the early stage but numerous pilots and partnerships are underway to rapidly bring this technology to market. In fact, as I had breakfast the other morning, the upcoming rollout of 5G licenses was on the front page of the China Daily newspaper.
Leveraging all of this infrastructure and generating massive amounts of data, is a population of over 1.42 billion people. On top of that, factories across China are fully leveraging the Industrial Internet-of-Things (IIoT) for intelligent automation and efficiency gains. All of these endpoints have created a data foundation in China that is unlocking new use-cases with artificial intelligence and business intelligence.
After a series of questions and answers, here are some additional takeaways based on our discussion on the big data landscape.
- The average salary of a data scientist in China is 1/4th of that of the same position and skill set in the United States. Many universities focused on producing data scientists, which has flooded the labor market in China with more supply than demand thus driving the salaries they can garner down. When asked what the most in-demand, high paying tech job was, Professor Hanfang explained it was an Algorithm Engineer.
- The government in China plays an important role in big data. Professor Hanfang explained that all government is big government in China. Local government in China is also big due to the political structure and density of development within cities (Over 160 Chinese cities have a population of over 1 million people, compared to only 14 in the US).
- China’s Ministry of Education supports lifelong learning opportunities that are as cheap as 10 RMB ($1.42 USD) to learn new specialized skills, such as big data, through online education most similar to MOOCs (Massive Open Only Courses) in the United States.