ChinAI #39: College Admissions and Industrial Upgrading
Welcome to the ChinAI Newsletter!
Hey, you know what would make a really great holiday present for that special someone in your life? Probably not a link to the ChinAI newsletter, but share it anyways!
These are Jeff Ding's weekly translations of writings on AI policy and strategy from Chinese thinkers. I'll also include general links to all things at the intersection of China and AI. Please share the subscription link if you think this stuff is cool. Here's an archive of all past issues. *Subscribers are welcome to share excerpts from these translations as long as my original translation is cited.
I'm a grad student at the University of Oxford where I'm the China lead for the Governance of AI Program, Future of Humanity Institute.
What can a College Admissions Advice Essay Tell us about China’s AI Scene?
This week’s main translation comes from our old friend Saidong - you may remember them as the writer of an epic strategy/rant/data-packed essay on how China can invigorate its chip industry, which we featured on an earlier ChinAI issue. This week we look at another epic piece on China’s higher education scene that can only be described as a Fivethirtyeight data journalism/Ask Amy advice column hybrid. Here’s the key question: how should students pick their undergrad university and their major?
Saidong stresses computer science as a major - it’s the key to social mobility, easiest for the working class to improve their situation (翻身) - he also argues that IT-related industries will still remain the fastest-growing ones as China’s overall growth slows.
Saidong analyzes the educational backgrounds of professors at various schools, highlighting the drastic upgrading in higher education (e.g. in 2005, nearly 40% of Peking University teachers did not have a doctorate). Beyond the usual suspects (Peking, Tsinghua, Shanghai Jiaotong, Fudan) schools like Shenzhen Southern University of Science and Technology and Suzhou University are leveraging their good geographical locations and investment in talent programs (attracting both top domestic and overseas scholars).
Some schools have extremely strong advantages in certain disciplines. For example, Harbin Institute of Technology’s SCIR laboratory is one of the very best in the field of natural language processing (NLP), producing: Baidu’s senior vice president Wang Haifeng, Tencent AI lab’s NLP leader Zhou Lianqiang, and Alibaba’s iDST NLP leader Lang Jun. Similar to students from countries, Saidong says that “for students in the popular directions of machine learning and artificial intelligence at Chinese universities with strong computer science departments in China, it is difficult to see even their shadows at the job market fairs for graduate students. Most of these students are well known early on, as renowned companies have “ordered them in advance.”
Here’s why it all matters in the big picture: “The ability to train PhDs in batches on a large scale shows that a country has the ability to industrialize, streamline, and standardize the training of scientific researchers, and has a large number of people who can collaborate with and exchange ideas in the scientific community.” Though not without its deficiencies — talent overly concentrated in certain areas and probably some overinflation of overseas degrees — China’s investment in higher education will be a huge factor in sustaining its AI dream. There’s a lot more good stuff in the full translation - I tried to add enough context and background about the gaokao college admissions process in areas where it may get confusing, but if not we can discuss it in the Google doc:
READ FULL TRANSLATION: COLLEGE ADMISSIONS IN THE ERA OF HIGHER EDUCATION TRANSFORMATION AND INDUSTRIAL UPGRADING
What Top Chinese Officials Are Hearing About AI Competition and Policy
Remember that special lecture we started translating a couple of issues ago? With a huge lift in editing from Graham Webster and the team at DigiChina, Cam Hickert and I finished up the full translation of a speech by Tan Tienu, deputy secretary-general of the Chinese Academy of Sciences, given to top officials at China’s 13th National People Congress.
Three quick takeaways: 1) It’s a impressively level-headed and relatively comprehensive speech that touches on a wide range of topics (the gap between narrow and general AI, the risks of an AI bubble in China, the need to control the impact of AI on society)
2) One reason why Tan thinks the “situation is heartening” for China’s AI development links back to this week’s feature translation: “In the past two years, Tsinghua University, Peking University, the University of the Chinese Academy of Sciences, Zhejiang University, Shanghai Jiaotong University, Nanjing University, and other universities have set up AI institutes.” Translation also presents some new publication metrics for mainstream AI academic conferences that show there’s a “huge gap between China and the United States in terms of basic strength at the frontier of AI.”
3) These types of documents are also interesting in that they show how Chinese leaders are trying to reframe AI competition. This section was particularly enlightening in that respect: “China is the only one in the lineup of developing countries that is expected to become a global leader in AI competition. It should adopt a route different from some countries’ ‘economic monopolization, technological protectionism, and trade bullying,’ and as soon as possible build up a structure of openly-shared, high-quality, low-cost, universally-beneficial, and global AI technology and application platforms, in line with the national ‘Belt and Road’ strategy.”
READ FULL TRANSLATION: SPECIAL LECTURE on AI
This Week's ChinAI Links
Chinese phrase of the Week: 一 趋之若鹜: (qu1zhi1ruo4wu4): an idiom sometimes used pejoratively, meaning to to “rush like ducks,” or scramble madly after something (often unobtainable), a wild goose chase.
Wired’s special issue on AI keeps hitting hard: last week recommended a profile on Fei-fei Li, this week it’s a feature on Karl Friston and his free energy principle, which the article claims is starting to spread into more mainstream deep-learning research. The passage that connected the dots a little for me discussed how a group at King’s College London that pitted 2 AI players against one another in the game Doom, and found that the agent driven by active inference (linked to free energy principle) performing better than one driven by reward-maximization.
Shout-out to What’s on Weibo — an independent news site run by Manya Koetse that explores trends on China’s social media. I’ve enjoyed recent pieces on social credit and the gene-edited baby twins case.
CSIS recently hosted a debate on China’s power, which had two great segments on Made in China 2025’s threat to global innovation (Scott Kennedy/Mu Rongping arguing for/against); China is likely to be the leader of the coming AI revolution (Edward Tse/Samm Sacks arguing for/against).
Chinese AI chipmaker Horizon Robotics is raising up to $1bn in a funding round, per Louise Lucas of Financial Times.
Thank you for reading and engaging.
Shout out to everyone who is commenting on the translations - idea is to build up a community of people interested in this stuff. You can contact me at email@example.com or on Twitter at @jjding99