Greetings from a world where…

ChinAI is now three years old, which means it should be able to walk in a straight line, but it rarely does. Thanks to all the readers, especially the 170+ paying subscribers who support ChinAI under a Guardian/Wikipedia-style tipping model. If you’ve been meaning to subscribe but just haven’t gotten around to it, no time like the present.

As always, the searchable archive of all past issues is here.

Reflections - Making Translation Newsworthy

What is newsworthy? This question should haunt everyone with a platform.

Last month, Stanford HAI published the AI Index Report 2021, a 222-page report on the state of AI, put together by an all-star team supported by a lot of data and strong connections to technical experts. What was newsworthy in this report? According to The Verge, “Artificial intelligence research continues to grow as China overtakes U.S. in AI journal citations.” In fact, the article takes its cue from what the report authors themselves deemed significant, given that “China overtakes the U.S. in AI journal citations” features as one of the report’s eight key takeaways.

Dig deeper into the data, however, and you’ll uncover alternative takeaways. Look at the cross-national statistics on average field-weighted citation impact (FWCI) of AI authors, for example, which gives a sense of the quality of the average AI publication from a region. Interestingly enough, the U.S. actually increased its relative lead in FWCI over China over the past couple years. According to the 2019 version of the AI Index, the FWCI of US publications was about 1.5 times greater than China’s; in 2021, that gap has widened to almost 2 times greater (p. 24).

So, working off the same materials as released in the AI index, here’s another way one could have distilled key takeaways: “The U.S. increases its lead over China in average impact of AI publications.” Or, if you wanted to be cheeky: “China lags behind Turkey in average impact of AI publications.” Just as newsworthy, in my opinion.

However, what I found most newsworthy about the AI Index went beyond horserace reporting about “who’s winning the AI race?!”Instead, I was most intrigued by the rise of commercially available machine translation (MT) systems, covered on page 64. According to data from Intento, a startup that assesses MT services, there are now 28 cloud MT systems with pre-trained models that are commercially available — an increase from just 8 in 2017. But wait … there’s more: Intento also reports an incredible spike in MT language coverage, with 16,000+ language pairs supported by at least one MT provider (slide 33 of Intento’s “State of Machine Translation” report).

Thanks to the State of MT report, we can also get a sense of the greater significance of translation advances. Let’s take Microsoft’s neural machine translation (NMT) service, priced at $10 per million symbols, for a ride. Recall that Microsoft achieved a milestone in 2018 for the first MT system reach human parity for translating news articles from Chinese to English. For $10 we can translate about 200,000 words, using an average of 4.79 symbols per word. How does one even grasp the potential of 200,000 words? That’s four times the number of Chinese characters in Wang Shuo’s great novella <<动物凶猛>>, one of so many Chinese language classics that have not yet been translated to English. These are, of course, very rough calculations — NMT works much better for news than novels, and the process would require a lot of post-editing — but they do open a window into the possibilities of NMT.

Somehow, these incredible advances in translation are not relevant to the effect of AI on U.S.-China relations, at least based on existing discussions. Compare the complete dearth of Twitter discussions centered on the following keywords: U.S., China, and “machine translation” against what you get when you replace “machine translation” with “facial recognition.” Consider another reference point, the recently published 756-page report by the National Security Commission on Artificial Intelligence (NSCAI). Sixty-two of those pages mention the word “weapon” at least once. Only nine pages mention the word “translation,” and most do not substantively discuss translation (e.g. the word appears in a bibliographic reference for a translated text).

Yet, I could make a convincing case that translation is more significant than targeting for U.S. national security. Think about the potential of improved translation capabilities for the intelligence community. Another obvious vector is the effect of translation on diplomacy. To the NSCAI’s credit, it notes how advances in MT could “transform the way we communicate across geographic and cultural barriers, enabling business, diplomacy, and free exchange of ideas” (p. 36). Consider also how MT could increase economic interdependence, which could have a pacifying effect on conflict escalation. Professor Steve Weber explores this in “The 2020s political economy of machine translation,” a working paper which analogizes MT to the container shipping revolution. Finally, recall the possibility of more translated books, which could also have security implications. After all, empirical studies have found that students have reduced threat perceptions of their host country after studying abroad, and what is reading, if not a study abroad experience.

So, how will AI advances affect U.S. national security and U.S.-China relations? Are the U.S. and China DECOUPLING in strategic technologies like AI? The maddeningly beautiful thing about studying AI + politics is that no one really knows. If you take translation as the key frame of reference for AI as opposed to weapons targeting or facial recognition, the answer to these questions is very different (#coupledforlife). This is made all the more complicated by the fact that AI is a general-purpose technology, with potential applications that far exceed translation, facial recognition, and weapons targeting. So, if you really wanted the true answer to the above questions, you would have to somehow predict the net impact of the distribution of all still-evolving AI applications.

Our answers to these questions and our determination of what is newsworthy about China’s development of AI, therefore, say more about us — our biases, our interests, our social influences — than they do about the Truth of the matter. I am no exception. Unlike journalists, translators do not have to pretend to be neutral. ChinAI is a product of my own biases, interests, and social influences.

That includes my fears. I vigorously pushback against overhyping China’s technological prowess in part because I fear how threat inflation could affect Asian Americans like me. I am quick to rebut claims of decoupling in part because I worry about what it means for my family’s ability to visit loved ones in China.

Newsworthiness doesn’t just have to be driven by our fears. And so, ChinAI is also a product of my dreams. Here’s one: One day, my kids will learn to read Wang Shuo’s 动物凶猛 in Spanish.

ChinAI Links (Four to Forward)

My four favorite issues of ChinAI from year 3 were all collaborations with contributors:

• ChinAI #98: Techlore - The Historical Rise of TSMC & Samsung in Semiconductors — Joy Dantong Ma found an epic poem of sorts about the history of Taiwan Semiconductor Manufacturing Company.

• ChinAI #104: Tencent 2020 AI White Paper — Caroline Meinhardt, a GovAI summer fellow, helped shed light on how Tencent is thinking about AI applications, including deepfakes.

• ChinAI #112: The Human Cost of 30-min Food Deliveries — Gabriele, San, and many other anonymous contributors helped translate a 15,000-word longform article on how drivers get trapped by food delivery system algorithms

• ChinAI #129: An Emotional Mess — Shazeda Ahmed guest-edited an issue about emotion recognition in China, featuring a translated article about backlash to emotion recognition applications in classrooms and her report (co-written with Vidushi Marda) on China’s emotion recognition landscape

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a Predoctoral Fellow at Stanford’s Center for International Security and Cooperation, sponsored by Stanford’s Institute for Human-Centered Artificial Intelligence.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a world where…

the Hawkeyes are still dancing

…Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the searchable archive of all past issues is here.

Feature Translation: China’s Tech Industry and Carbon Neutrality

Context: Last week, Leiphone published a fascinating piece on state-of-play re: China’s tech industry and the momentum toward carbon neutrality (碳中和). Many Chinese media have proclaimed 2021 to be the “first year of the carbon neutral era;” the government work report released at the Two Sessions included carbon neutrality for the first time.

Key Takeaways:

• According to a reports by Tianfeng Securities Research Institute and Greenpeace, the tech industry’s carbon emissions have continued to rise in recent years. Specifically, it ranks high in terms of Scope 3 emissions, which are indirect emissions that occur in a company’s value chain. By 2040, the information communications technology (ICT) industry is expected to account for more than 14 percent of global greenhouse gas (ghg) emissions.

• Data centers were responsible for 87% of Tencent’s carbon emissions, and as a whole, China’s data centers consumed 2% of the country’s total electricity usage in 2018.

• After reviewing commitments by U.S. tech giants, the piece states that Chinese tech companies lag behind in terms of clear timetables to reach carbon neutrality, though some have made notable moves. These include: Ant Group’s recently announced goal of net-zero emissions by 2030, Alibaba and Tencent’s use of liquid cooling tech to reduce heat emissions in data centers, and Baidu’s use of photovoltaic cells to power its data centers.

• Given the significant environmental costs of training large AI models, this is definitely an area to track for the future. Announcements are nice, but transparent, verifiable action is better. As of a January 2021 Greenpeace report (in Mandarin), Chindata Group (秦淮数据) was the only Chinese firm that had committed to achieving 100% renewable energy by an end date.

ChinAI Links (Four to Forward)

Should-read: Asian Americans at State Department confront discrimination

Ryan Heath for Politico on assignment restrictions for Asian American diplomats:

One former diplomat subject to restrictions was Rep. Andy Kim (D-N.J.), a Korean-American born in Boston who told MSNBC Wednesday that even though he had ‘top secret security clearance’ and had served in Afghanistan, “One day I was told by the State Department that I was banned from working on anything related to the Korean Peninsula.”

Kim said he was shocked because he had never applied to work on any issues related to the Korean Peninsula. He labeled the decision xenophobic and said that what hurt most was “this feeling that my country didn’t trust me.”

A statement signed by over 100 Asian Americans working in national security and diplomacy, argued Thursday that the increasing U.S. focus on competition with China, has exacerbated “discrimination, and blatant accusations of disloyalty simply because of the way we look.”

Related: see Amy Chang’s Twitter thread sharing her personal experience as an Asian American women working on U.S.-China relations:

Amy Chang @quelquefois

Bear with me as I am still processing the numerous social forces and identities that have culminated in this thread about being an Asian American female working in security and the movement around #StopAsianHate #StopAAPIHate 1/

March 20th 2021

31 Retweets132 Likes

Should-read: Key Concepts in AI Safety

The first in CSET’s series on “AI safety” by Tim Rudner and Helen Toner, which goes over how to develop machine learning systems in a safe and reliable way. They introduce some key concepts, including specification, which refers to “defining a system’s goal in a way that ensures its behavior aligns with the human operator’s intentions.”

Should-read: The AI Wolf Refuses to Play the Game (Mandarin)

Speaking of misspecified AI systems, a funny (and slightly disturbing) example that went viral last week, as reported by xinzhiyuan (AI Era): researchers had set up a wolf vs. sheep game. Instead of trying to eat as many sheep as possible, which was the intent of the game, the wolf chose to run into a rock and kill itself because that somehow gave more points than not catching any sheep.

Should-read: China sours on facial recognition tech

For Protocol, Zeyi Yang covers a 10-minute investigative segment aired on the CCTV annual consumer rights gala, which “revealed that facial recognition security cameras located at chain stores nationwide have been picking up shoppers' personal information without their knowledge or consent. The revelations ignited a furious backlash against the companies. . . It's another instance of grassroots pushback against surveillance tech in China, a global leader in surveillance research as well as in deployment. The central irony went unremarked: that Beijing has become both the critic and perpetrator of mass surveillance.”

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a world where…

vision is always peripheral

…Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the searchable archive of all past issues is here.

Feature Translations: Does anybody want to be China’s first computer vision stock?

Come, gather around the fire, and allow me to spin you a tale of China’s “Four Computer Vision Dragons” — CloudWalk, Megvii, Sensetime, and Yitu — and their quest to become China’s first computer vision stock (AI视觉第一股).

Our story begins with Megvii, founded back in 2011 in Beijing. Enduring AI’s cold winter in its early years, Megvii filed an IPO prospectus in August 2019 to be listed on the Hong Kong exchange, with an expectation of raising USD$500m-$1bn. By then, Megvii claimed the largest share of the Chinese market in cloud-based identity authentication solutions enabled by facial recognition.

Two months later, the U.S. government blacklisted Megvii (along with Sensetime, Yitu, and other companies) for its alleged involvement with human rights abuses in Xinjiang. In February of 2020, Megvii let its IPO application lapse. Last week, Megvii filed for an IPO on the Shanghai Stock Exchange’s STAR market (China’s NASDA-esque index aimed to encourage domestic investment in Chinese tech companies). See Shen Lu’s excellent protocol piece for more on Megvii’s updated IPO documents.

The giant in the room, Sensetime, was founded much later in 2014 in Hong Kong, and has earned the label of “The World’s Most Valuable AI Startup.” In 2017, SenseTime’s founder told Reuters that it would consider a US, Hong Kong, or mainland listing, but now it is mulling a dual listing in Hong Kong and on China’s STAR market.

Next up is CloudWalk, founded in 2015 and headquartered in Guangzhou. Curiously, the U.S. government did not list CloudWalk in the initial entity list in October 2019, though CloudWalk was added later in May 2020. I’ve previously highlighted that CloudWalk is unlike the other three dragons because of its financing from Chinese government funds, making it closest to a “purely domestically funded enterprise” (纯内资) . In December 2020, CloudWalk also filed for an IPO on the STAR Market.

Et tu, Yitu? What about the remaining dragon? Founded in Shanghai in 2012, Yitu has been around for a while. Some of its early work included work with the Suzhou Public Security Bureau to identify cars with fake license plates. In November 2020, it also filed for an IPO — guess where? — on the STAR Market, aiming to raise more than 7.5 billion RMB, and to become the first of China’s AI unicorns to go public. But then, Yitu suspended its IPO application last Thursday, March 11. Why? Our tale continues with the help of two articles by AI科技评论(aitechtalk), which provides in-depth reports on developments in the AI industry and academia, and 全景财经 (quanjing caijing), a company that analyzes China’s capital market.

According to aitechtalk, Yitu’s filing “provoked controversy”: the STAR market raised a total of 47 issues in its initial comments on the prospectus, including questions about Yitu’s family trust and special voting rights structure. Aitechtalk digs deeper into the prospectus:

• Yitu claims 800 government and corporate end customers across basically all Chinese provinces and more than 10 overseas countries and regions.

• As one would expect of a tech-driven company, R&D expenses accounted for 147%, 96%, 92% and 100% of revenue for the previous four reporting periods.

• What really surprised me was Yitu’s commitment to System on Chip (SoC) projects. Here’s how one Equalocean analysis described Yitu’s pivot from software-based solutions to compute-based solutions like its Questcore chip: “With the AI chip launch, Yitu's revenue structure soon changed. The once-dominant software business was squeezed while the integrated solutions (software and hardware combined) saw an increasing weight.”

This week’s second article, from quanjing caijing, provides a more panoramic view. Yitu is one of 37 companies that have suspended their IPO applications in the Shenzhen and Shanghai stock markets. A total of 152 companies have terminated their IPO status. Hesai Technology, for instance, presumed to be the "first Lidar stock" also withdrew its IPO application from STAR this week. The China Securities Regulatory Commission is tightening up rules and scrutiny for startups aiming to go public on the STAR Market. The article calls it a tide of listing withdrawals (撤单潮).

As for which one of the dragons will ride/fly out the tide, that’s a story for another day.

ChinAI Links (Four to Forward)

Must-read: You’re Doing It Wrong — Notes on Criticism and Technology Hype

Lee Vinsel, a professor in science, technology, and society at Virginia Tech, published a Medium post last month that is really really worth your time. In the piece he calls out criti-hype:

Recently, however, I’ve become increasingly aware of critical writing that is parasitic upon and even inflates hype. The media landscape is full of dramatic claims — many of which come from entrepreneurs, startup PR offices, and other boosters — about how technologies, such as “AI,” self-driving cars, genetic engineering, the “sharing economy,” blockchain, and cryptocurrencies, will lead to massive societal shifts in the near-future. These boosters — Elon Musk comes to mind — naturally tend to accentuate positive benefits. The kinds of critics that I am talking about invert boosters’ messages — they retain the picture of extraordinary change but focus instead on negative problems and risks. It’s as if they take press releases from startups and cover them with hellscapes.

At their most ridiculous, hype-filled criticisms become what historian David C. Brock calls “wishful worries,” that is, “problems that it would be nice to have, in contrast to the actual agonies of the present.” . . . Part of Brock’s point is that wishful worries are a kind of entertainment. We are, after all, a people that regularly feasts upon dystopian science fiction. Imaginary fears can be fun.

The rest of the piece argues that Shoshana Zuboff’s book, The Age of Surveillance Capitalism, overstates the abilities of social media firms to directly influence our thoughts, and also supplies “a preliminary history of how criti-hype became an academic business model by taking a look at the examples of the Human Genome Project, nanotechnology, ‘AI,’ and a few others.”

Must-read: The US Is Building Walls Around Science, and We’re All Poorer for It

Yangyang Cheng’s essential writing for Vice on the case of Gang Chen, a MIT professor accused of failing to disclose contracts and appointments from Chinese entities. It’s the first piece I’ve read on Gang’s case and the DOJ’s misguided China Initiative that actually treats Gang like a human being. But it also does so much more: expertly deconstructs the Justice Department’s misleading use of an email excerpt, provides important historical context, and reminds us that the Biden administration has not ended the China Initiative and has kept on Trump-appointed FBI Director Christopher Wray, who once declared China a threat that requires a “whole-of-society” response.

Should-read: Understanding Chinese Government Guidance Funds

A really impressive CSET analysis on government guidance funds (GGFs) by Ngor Luong, Zachary Arnold, and Ben Murphy. What really stands out is the appendix from page 33-63 which provides key translated excerpts from all the Chinese-language sources the report is based on. Two main findings:

• GGFs “fail to live up to their ambitions, weakened by unrealistic goals, bureaucratic constraints, incompetent management, risk aversion, and a lack of market discipline.”

• However, “guidance funds still have advantages over China’s traditional industrial policy mechanisms. And today, a subset of disciplined, market-oriented guidance funds is successfully raising money and investing in projects.”

Should-read: How Standard Setting Can Help Taiwan Grow Its Global Role

For the Carnegie Endowment for International Peace, Evan A. Feigenbaum and Michael Nelson tackle the important question of Taiwan’s approach to setting international technical standards for strategic technologies.

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a world where…

minari grows where it’s planted

…Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the searchable archive of all past issues is here.

Feature Translation: Chinese cyberstrategy discussions reference weaponized interdependence

In July 2019 Henry Farrell and Abe Newman published “Weaponized Interdependence: How Global Economic Networks Shape State Coercion” in International Security, a leading international relations journal. Here’s the quick summary:

Liberals claim that globalization has led to fragmentation and decentralized networks of power relations. This does not explain how states increasingly “weaponize interdependence” by leveraging global networks of informational and financial exchange for strategic advantage . . . Highly asymmetric networks allow states with (1) effective jurisdiction over the central economic nodes and (2) appropriate domestic institutions and norms to weaponize these structural advantages for coercive ends. In particular, two mechanisms can be identified. First, states can employ the “panopticon effect” to gather strategically valuable information. Second, they can employ the “chokepoint effect” to deny network access to adversaries. Tests of the plausibility of these arguments across two extended case studies that provide variation both in the extent of U.S. jurisdiction and in the presence of domestic institutions—the SWIFT financial messaging system [which was used to reinforce U.S. and European sanctions against Iran] and the internet—confirm the framework's expectations.

The article clearly tapped into something that has resonated with a lot of people thinking about the intersection between economics and security. It’s the 5th most read article from the journal in the past year. In this week’s feature translation, I look at how one Chinese scholar uses the weaponized interdependence framework to think about international technological competition.

Context: China Information Security is an influential Chinese-language journal that covers topics like cybersecurity and internet governance. In the journal’s first issue of 2021, Xu Xiujun (PhD in int’l relations, research fellow at the Chinese Academy of Social Sciences) published an interesting essay on network security governance in the “Cyberspace Strategy Forum” vertical. This section of the journal often contains big-picture thinking, such as Dr. Hong Yanqing’s piece 2018 article (links to English translation) on “Designing Risk-based Critical Information Infrastructure Protection In China.”

In a section on how information technology could reshape the landscape of international competition, Dr. Xu essentially restates the thesis of Farrell and Newman’s article on weaponized interdependence. Note how he references the same two mechanisms (panopticon/oversight and chokehold/blocking) as well as the SWIFT messaging example:

“Many scholars once believed that economic globalization has made information more dispersed and the world more flat. However, as globalized economic activities and information exchanges have increased, some "central nodes" where information gathers have become increasingly prominent, and the countries occupying these nodes therefore have ‘oversight powers’ [监视权] and "blocking powers" [阻断权] that restrict the behavior of other countries. With the increasing number of Internet users, the control of global Internet-related business activities such as cross-border payment and e-commerce is increasingly concentrated in the hands of a few countries . . . Moreover, this asymmetrical interdependence can also become a weapon for sanctions against other countries. For example, in the global financial system, the SWIFT system has this effect. Since the establishment of the SWIFT system by some European and American banks in 1973, more than 200 countries and 11,000 financial institutions have used this system. SWIFT is equivalent to the information aggregation center for global financial transactions and plays a key central node role in the global financial system. This makes it possible for countries that control this system not only to track transaction records between countries and institutions, but also to impose sanctions on specific countries and institutions by stopping the provision of financial information services.”

How should China adapt to a world of weaponized interdependence? Dr. Xu emphasizes the need to enhance China’s independent innovation capability in technologies like AI. He also highlights the importance of formulating international technical standards, noting that China accounts for less than 1% of ISO standards. He concludes:

“Only by occupying an advantageous position of interdependence in the network domain can we effectively prevent China from becoming a target of weaponized interdependence [相互依存武器]. Unlike Iran and other countries, relying on existing technologies, China has the ability to actively increase its own central nodes and network control in certain areas, and strengthen its ability to counterattack the United States with weaponized interdependence. To this end, China should rely on major breakthroughs in the development and application of core technologies in the field of network information such as high-performance computing, mobile communications, quantum communications, Beidou navigation (satellite system), core chips, operating systems, etc., to continuously improve network control and form a powerful deterrent to prevent other countries from abusing weaponized networks to generate security threats.”

FULL TRANSLATION: The international environment and countermeasures of network governance during the "14th Five-Year Plan" period

ChinAI Links (Four to Forward)

Must-read: In-depth Analysis of Independent and Controllable Chips (Mandarin)

A piece by Suny Li for SiP与先进封装技术 that analyzes the chip supply chain across nine segments, three in each of the three main phases of chip development:

• chip design (EDA, IP, and design) ->

• chip manufacturing (equipment, processing, and materials) ->

• packaging and testing (packaging design, product packaging, and chip testing)

The author compiles the top ten players in each of the nine segments. For example, below is the table for IP in chip design. The only mainland Chinese firm, Verisilicon, has 1.8% of the market share.

Should-read: How global tech executives view U.S.-China tech competition

Featured in Brooking’s TechStream, Christopher A. Thomas and Xander Wu conducted a very interesting survey of 158 senior business executives working for American, Chinese, European, Japanese, Taiwanese, and Korean global high-tech firms whom they polled about the impact of U.S-China tensions on their industry. Here’s one fascinating takeaway:

“Tensions between the United States and China pose difficult questions for tech companies headquartered outside the two countries. But based on our survey results, these firms do not plan on choosing sides. As Figure 11 illustrates, an overwhelming 98% of respondents indicate that these companies from third-party countries will align with neither China nor the United States. Two out of three respondents expect companies based outside the United States and China to play the two countries off one another to get benefits from each. As Figure 10 above shows, two out of three also said they would further invest in China or localize operations there, despite more than 95% believing multinational companies in China will face an uneven playing field there.”

Should-read: 2021 AI Index by Stanford HAI

Fourth edition of the most comprehensive report on AI measurement is out. One of the takeaways that has received a lot of media attention is: “China overtakes the US in AI journal citations.” This comes after China surpassed the U.S. in total journal publications. I don’t see this as an especially noteworthy indicator of the quality of AI research, since you can rack up more citations by publishing large amounts of derivative research. That’s why the index shows that China lags far behind the U.S. and the European Union in field-weighted citation impact of AI publications as well as conference publications, which are more relevant than journal articles for AI. For more, see my written testimony before U.S.-China Economic and Security Review Commission on why China is far from an AI superpower poised to overtake the U.S. in this domain.

Should-read: National Security Commission Final Report on Artificial Intelligence

Speaking of overhyping China’s AI capabilities…here’s p. 20 of this newly released report: “If the United States does not act, it will likely lose its leadership position in AI to China in the next decade.” But if you make it past the bluster in the beginning — or take it for what it is: obligatory marketing to cater to a DC audience hooked on a narrow vision of national security — there’s some smart moderate policy ideas in the report (e.g. chapter 7 on establishing justified confidence in AI systems).

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a world where…

kingdoms are transcendent

…Please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the searchable archive of all past issues is here.

Feature Translation: Sensetime’s Open Source Strategy

This week, we’re taking a look at an August 2020 article by the China Software Developer Network (CSDN), the largest software developer community in China, which also provides IT news coverage.

Context: Longtime readers will recall a white paper on China’s development of AI open source software, which we translated in ChinAI #22. As the white paper emphasized, there is a push for “independent and controllable” (自主可控) AI software, as the main open-source deep-learning frameworks are developed by U.S. companies (Facebook’s PyTorch and Google’s TensorFlow).

At the start of 2020, Megvii and Huawei open-sourced their respective deep learning frameworks: MegEngine and MindSpore. Now everyone’s attention is on Sensetime, the AI unicorn with 4,000 employees. What will it do with its independently developed deep learning platform — SenseParrots? To find out, CSDN interviewed Prof. Dahua Lin, a co-founder of Sensetime.

Key Takeaways: Sensetime’s open source strategy is different from those of Megvii and Huawei. Instead of open-sourcing SenseParrots, its bottom-layer deep learning framework, it has chosen to first open-source its OpenMMLab, which includes upper-level algorithms and application platforms in deep learning for computer vision. The OpenMMLab Github repository has racked up 20k stars on Github (for reference, TensorFlow has 154k, and MegEngine has 3.8k)

• Why OpenMMLab and not SenseParrots? PyTorch and TensorFlow have consolidated huge network effects and first-mover advantages at the bottom layer. Thousands of papers and products are built on these two frameworks, and these are rich ecosystems. Simply open-sourcing a training framework is not enough to compete; PyTorch and TensorFlow support multiple layers of algorithms, tools, and various engineering environments. Instead, Sensetime chose to compete at the layer of algorithm and engineering development for building upper-level applications in deep learning for computer vision.

• What’s the grand strategy? Dr. Lin claims, “Once we have ecosystem-level influence, we can use this as an entry point to plan the next step. In the future, we will open the lower-level deep learning framework SenseParrots at the right time.” The idea is that once many upper-level applications are built with OpenMMLab, then the migration costs for developers will be greatly reduced once Sensetime open-sources the SenseParrots infrastructure.

• Other interesting tidbits: Dr. Lin discusses the main advantages of OpenMMLab, his hopes for something like the generative pretrained transformer (GPT-3) in the computer vision field, and his thoughts on the future division of labor in an AI ecosystem built on large models like GPT-3.

See FULL TRANSLATION: 20,000+ stars on Github, a domestically developed AI open source software starts to breakthrough | interview with Lin Dahua, Sensetime cofounder

ChinAI Links (Four to Forward)

Must-read: Chinese tech giants can change but the state is still their number one stakeholder

From the Ranking Digital Rights (RDR) project, Rebecca MacKinnon writes: “Chinese companies have taken meaningful steps over the past five years to protect consumer privacy and security from threats that are unrelated to Chinese government surveillance . . . Their progress shows that Chinese companies can and do respond to pressure from foreign governments, investors, and even users. But the extent to which they can improve policies and practices affecting users’ human rights is severely handicapped by China’s legal and political system. The hard reality is that Chinese companies are powerless to protect users (whether they are in China or abroad) from digital rights violations by one of the most powerful—and unaccountable—governments in the world.

Should-attend: Audrey Tang on Taiwan’s digital democracy, collaborative civic technologies, and beneficial information flows

The next session in GovAI’s webinar series will go live on Tuesday, March 9th (1000-1130 GMT). Audrey Tang, Taiwan’s first Digital Minister will discuss collaborative civic technologies in Taiwan, their potential to improve governance and beneficial information flows, with Hélène Landemore, an Associate Professor at Yale, as a discussant.

Should-read: Citi Can’t Have Its $900 Million Back

An absolutely unreal “horror story about software design.” Citibank accidentally wired $900 million to some hedge funds last August. Some did not send the money back. Citi sued them and recently lost the suit. Matt Levine read the judicial opinion and wrote about the case for Bloomberg. What shocked me was the rigidity of the software that Citi uses to process its financial translations (Oracle’s Flexcube). I think for every article about the newest groundbreaking advance in deep reinforcement learning, we need 10 articles about just basic software development.

Should-read: How a con man defrauded billions out of a Wuhan chip project

China Money Network translates a detailed report by Chinese media 36kr on the scam that was Wuhan Hongxin Semiconductor Manufacturing Corporation . A tale on the perils of China’s industrial strategy on chips — h/t to ChinAI subscriber Ben Mueller for pointing me to the story. For related coverage, see this Caixin article. Careful readers of ChinAI will recall that we covered Hongxin’s troubles last November.

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99