Welcome to the ChinAI Newsletter!

Quick correction on last week’s list of top ten queried items for waste sorting:

1. dog poop wrapped in a napkin
2. dog poop wrapped in a sack
3. cat poop wrapped in a sack
4. dog poop wrapped in a newspaper
5. condoms
6. takeaway bags
7. bubble tea cups
8. wet wipes (h/t to Ryan Soh, ChinAI contributor, for correcting his own correction — we had 湿厕纸 translated as soiled toilet paper in last week’s issue)
9. cray fish (only one labeled as wet waste)
10. nose booger wrapped in a napkin

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

Feature Translation: A Research Outline for AI in China’s Education Field

This week’s translation and analysis are brought to you by Kwan Yee Ng, a researcher at the Center for Long-term Priorities. Written by a group of academics and researchers (including two professors at Huazhong Normal University AKA Central China Normal University), the piece is a meta-analysis of 176 Chinese academic papers on AI in education published from 2017 onwards and a critique of the quality of this output. Analysis by Kwan follows:

The increasing volume of Chinese papers on AI has been held by some as another piece of evidence that China is set to lead the world in AI (see here, here, and here). This article should serve to unpack this trend for further scrutiny. While the article explicitly pertains to Chinese research on the intersection between AI and education, as someone who's been doing their fair share of trawling Chinese academic platforms to research Artificial General Intelligence (AGI), I think the findings also apply to Chinese academic papers on AI more generally. Some takeaways:

• The dearth of referencing and frequency of cross-references: as per the article, there is a “network of cross-referential research” and “in 2017 an all-round flourishing [in the volume of publications] began. It is important to point out that between 2016 and 2017, the volume of publications increased sharply but the number of references dropped significantly. This is worthy of vigilance because the relative lack of references means that the research is less objective and scientific.” 

• A lot of papers read more like op-eds and do not suggest an informed understanding of AI technology: the article laments "irregular academic terminology" and finds that "from the perspective of the whole, there is more qualitative research and less quantitative research, more normative research and less empirical research."

• A lack of breadth in case studies, among other things, indicate the strength of influence these cases have on the collective memory: “Most examples or quotes are limited to a few "star" cases, such as AlphaGo, autonomous driving, IBM Watson, GoogleBrain cat face recognition, and ImageNet image recognition competitions.”

Dig Deeper: The two lead authors, Liu Kai and Hu Xiangen, are notable proponents of NARS, Non-Axiomatic Logic Reasoning System, an AGI model theorised by Wang Pei. More on NARS here, and Wang’s description of NARS:

• “The design of NARS is based on the belief that the essence of intelligence is the principle of adapting to the environment while working with insufficient knowledge and resources. Accordingly, an intelligent system should rely on finite processing capacity, work in real time, open to unexpected tasks, and learn from experience.”

FULL TRANSLATION: A Research Outline for AI in China’s Education Field

ChinAI Links (Four to Forward)

Should Read: China Due to Introduce Face Scans for Real-Name ID when Registering a New Mobile Contract

The Ministry of Industry and Information Technology announced regulations in September for telecom operators to verify people’s real-name ID with facial recognition when they get a new sim card for their phone. I provided some comments on how there has been more pushback to China’s widespread adoption of facial recognition technology. Next week’s translation will highlight some of this pushback.

Should Read: Automation Impacts on China’s Polarized Job Market

Arxiv preprint by Chen et al. (researchers from Commonwealth Scientific and Industrial Research Organisation, Monash University, Sun Yat-sen University, and MIT Media Lab) — h/t to Remco Zwetsloot for sharing this.

Summary: "China’s top-down, centrally planned specialization of cities makes large Chinese cities less resilient to impact from automation technologies," as compared to large U.S. cities which are more resilient to the impacts of automation because of a more diversified job market. Related analysis by Frey et al. 2016 of Oxford Martin School found that 47% of US jobs were at risk of computerization compared to 77% in China. The authors state that well-known large cities in China, such as Beijing, Shanghai, Guangzhou, and Shenzhen, exhibit resilience to automation technologies, whereas other large “specialty cities” such as Nanyang which specializes in farming are more susceptible to the impacts of automation.

*I think it’s a useful line of inquiry but not convinced by their method. On a quick scan of the paper I didn’t find how they proxied city size. The authors claim that Nanyang (prefecture level city in Henan province) is “the fifth largest city in China” but that doesn’t square with existing lists of largest cities. I do think there’s a huge gap in our knowledge about cities like Nanyang and Zhumadian compared to the more well-known large cities of China.

Two Links on the We Can’t Break Up Big Tech Companies Because We Need to Compete with China Narrative:

This narrative got very high-profile coverage when Zuckerberg’s leaked notes for his Congressional testimony mentioned it as a key talking point. Zuckerberg also repeated this argument in hearings about its digital currency project in October. Two well-argued, opposing views on the issue follow, looking at Facebook and Qualcomm:

• Two researchers and social entrepreneurs argue that breaking up Facebook could open paths for Chinese-led alternatives that are more prone to surveillance, with a focus on the situation in Myanmar.

• The Qualcomm antitrust case also features the Too Big and Necessary to Compete with China narrative. Matt Stoller breaks down why more robust antitrust may actually help U.S. companies compete better.

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 Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.

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

Welcome to the ChinAI Newsletter!

At ChinAI we’re trying to find the treasure amidst the trash, and I especially like stories that look at what affects people in their daily lives, so this week’s translation was especially fun to prepare. This past July, Shanghai implemented a regulation that makes household waste sorting compulsory, 22,000 tons of daily household waste must be sorted into four buckets: dry, wet, recyclable, and hazardous. SCMP’s Inside China podcast did an amazing episode that gives great context on this. Many apps have popped up including one that lets users query which bin certain items should be sorted into:

The ranking of the ten most queried items:
1. dog poop wrapped in a napkin
2. dog poop wrapped in a sack
3. cat poop wrapped in a sack
4. dog poop wrapped in a newspaper
5. condoms
6. takeaway bags
7. bubble tea cups
8. soiled toilet paper (h/t to Ryan Soh, past ChinAI contributor for correcting this -- I had it as heated toilet paper before)
9. cray fish (only one labeled as wet waste)
10. nose booger wrapped in a napkin

*More takeaways in the feature translation section

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

Full Translation: 46 Chinese Cities Will Invest 20 billion (RMB) to promote waste sorting. Where are the opportunities for AI companies?

A cool longform piece by Wei Pang and Tai Lang of jiqizhixin on the opportunities for AI companies amidst China’s push to have 46 major cities (including Shanghai) recycle 35 percent of waste by 2020. What I like about this piece is it takes the concept of AI + recycling and breaks it down into the front-end, medium, and back-end process of a city’s waste management process and then analyzes where AI may play a role in each. The piece’s finding of middling returns to opportunities for AI applications matches with the SCMP’s Inside China podcast take on narratives of hi-tech solutions, eco-dictatorships, and social credit punishments for incorrect trash sorting.

Key Takeaways

Why is there such an imperative for more waste sorting in Shanghai?

1) Land is getting more and more expensive and landfill storage (where 70% of Shanghai’s waste goes) takes up a lot of land

2) The landfill capacity squeeze has caused some high-profile scandals (e.g. the Taihu garbage dumping incident in 2016 when more than 20,000 tons of construction waste were transported from Shanghai and secretly dumped on an island in a neighboring province)

3) Shanghai is very different from other large cities in Europe and the U.S. since the proportion of household waste in SH is much higher. It’s very similar to Tokyo in many respects except only 3% of household waste in Tokyo went to landfills, and incineration as a method for waste reduction was as high as 75%. But NIMBY tensions have complicated the building of waste incineration plants in China, so the authors make a final comparison between Shanghai’s approach and Taiwan’s experience where landfills are also overwhelmed and waste incineration plants are also continuously being protested. Taiwan has focused on front-end reductions of waste, such as waste sorting, recycling, and reuse.

The numbers breakdown of why sorting is important and where companies are trying to carve out a niche

• Front-end recycling process: Some rough calculations in Beijing context — for one ton of waste if there is no sorting at the frontend, the processing cost is about 2,000 RMB. With sorting, the cost can be reduced by about half or more. Some companies have designed essentially delivery apps for waste, but the labor costs are relatively high (6RMB to collect at the door when average customer’s waste is on average only worth 4RMB). Some companies have just put smart trash cans in the community, so residents who deliver their waste receive rewards from the platform, but many smart waste bins have already gone bankrupt. If AI is used for these smart bins, it’s mainly used for identity authentication via face scan. The big players are also involved. Alibaba’s e-commerce site Taobao has an AI waste recognition function. Alibaba, Xiaomi, and Jingdong have invested in electronic waste recycling companies.

• Middle of recycling process: after waste is collected it has to be transported to a specialized sorting center. According to 2015 data, Beijing produces more than 20,000 tons of waste a day, and the daily costs of transporting this waste is three or four hundred million. In the authors’ view, there have been fewer setbacks in this segment, as many companies are working in the smart dispatch market to reduce waste transportation costs, though they note more prominent examples such as US company Rubicon Global and Swedish efforts in waste management.

• Back-end of recycling process: at the waste processing plant, workers do labor-intensive, hard (150 RMB to go through 150-200kg of waste per day and find recyclables mixed in with the solid waste or filter out impurities not conducive to incineration), smelly work to sort waste. While this seems like an ideal application case for robots, interviews by the authors reveal that many robot startups have given up and most of the use cases come from abroad (e.g. Baidu Ventures invested in an American startup that uses machine vision and robotic arms for accurate waste identification and sorting)

DISCUSSED IN THE FULL TRANSLATION (in the style of Believer magazine): the Uber of waste transportation, a company with the motto “trash is cash,” garbage used as cement kiln fuel

FULL TRANSLATION: What are the opportunities for AI companies in waste sorting?

ChinAI Links (Four to Forward)

Must Read: Exposed: China’s Operating Manuals for Mass Internment and Arrest by Algorithm by the International Consortium of Investigative Journalists

More than 75 journalists from the International Consortium of Investigative Journalists and 17 media partner organizations have revealed in the Chinese government’s own words how mass detention camps in Xinjiang are run. The leaked classified Chinese government documents include a November 2017 operations manual with detailed guidelines for managing the camps and four shorter intelligence briefings (“bulletins” for managing the Integrated Joint Operation Platform [IJOP], a mass-surveillance and predictive-policing program that analyzes data from Xinjiang)

Key points (Reported in Bethany Allen-Ebrahimian’s piece “Exposed”) — later notes will draw from other pieces collected in link above:

1. This is the largest mass internment of an ethnic-religious minority since WWII - “Bulletin No. 14” notes that in a seven-day period in June 2017, security officials rounded up 15,683 Xinjiang residents flagged by IJOP and placed them in internment camps (in addition to 706 formally arrested).

2. The manual reflects the dehumanizing treatment of inmates in the camps: despite instructions to ensure health and safety, eyewitness accounts have testified to torture, beatings, and death of prisoners. The manual reveals a points-based system that classifies inmates by degree of security required and rewards them for ideological transformation and compliance.

3. The ICIJ investigation emphasizes the government’s “artificial-intelligence-powered policing platform” (IJOP): this system amasses vast amounts of personal data via warrantless manual searches, facial recognition cameras, app usage to identify those who need to be detained.

Longer Reflections on how the role of AI in this story. I want to make clear that this is in no way a criticism of the incredible reporting by ICIJ nor should it detract from the main thrust of the piece or the notes above. This is more my personal reflection on what the significance of AI is to a story like this. What I’m trying to get my head around is the degree to which it is useful to characterize IJOP as an AI-powered policing platform as well as the degree to which drawing distinctions on these issues even matters.

• James Mulvenon, who directs an intelligence contractor for several U.S. agencies, describes IJOP as “machine-learning, artificial intelligence, command and control” platform that substitutes artificial intelligence for human judgment and a “cybernetic brain” central to China’s most advanced police and military strategies.

• But what does this “cybernetic brain” consist of? Allen-Ebrahimian writes, “IJOP then uses an as-yet-unknown algorithm to create lists of people deemed suspicious.” Based on my read of the four bulletins and other reporting, I would describe IJOP as a big data platform that probably uses very basic algorithms that you could perform on an excel sheet (e.g. counting up the number of red flags, such as foreign nationality or use of certain apps, to identify targets). It might even be a stretch to think that logistic regression is even being used let alone deep learning-based algorithms.

• Don’t get me wrong. This type of granular data collection is concerning. Scilla Alecci’s piece highlights how users of the mobile file-sharing application called Zapya, which the Chinese gov. has flagged as a vehicle to disseminate religious material, have been targeted for further investigation. And the Human Rights Watch’s original report on the IJOP system highlights a range of other sources the platform collects data from (facial-recognition through CCTV cameras, wifi sniffers to collect identifying addresses of computers and smartphones).

• I just want to tug on what exactly this "unknown algorithm” is, and why we are calling it AI. It ties into some of my previous posts about the AI abstraction problem and whether AI has become a meaningless concept like ICT. The use of facial-recognition (pretty clear application of AI) to identify people is one component of data collection, but that doesn’t amount to an AI brain that produces lists of people deemed suspicious. Anyways these semantics don’t mean that much to the people facing persecution and separation from their families, and that should be the main takeaway here. But I think there’s something useful to pull out here about how we assign causal power to technologies and whether differentiating between the specific types of technologies that are enabling a certain situation to occur matters at all.

Should Read: Hong Kong Holds a Protest Election but Will it Produce a Protest Result? by Suzanne Pepper

The answer was a resounding yes as pro-democracy candidates now control 17 of the 18 District Councils. Suzanne Pepper’s piece on the eve of the election gives excellent context for:

• the function of the District Councils (neighborhood-level problem solving and implementation of gov policies)

• why they matter (feed into more democratic influence in choosing Chief Executive and seats in Legislative Council)

• why pro-Beijing candidates had dominated the councils before the recent election (better organization and splintering of pro-democracy forces).

Should Listen: four pods on China’s recycling revolution via SCMP’s Inside China podcast

Very relevant to this week’s translation: Richard Brubaker, long-term Shanghai resident and founder and managing director of NGO Collective Responsibility, talks about his company’s work to map the pre-existing network of informal recyclers and trash pickers, and how the new four bin system has affected that.

Brubaker also wades into media reports dominated by stories of hi-tech surveillance and social credit punishments for incorrect trash sorting and talks of the day-to-day realities for Shanghai residents he’s witnessed.

Should Read: Winners and losers in US-China scientific research collaborations

By Jenny J. Lee and John P. Haupt from the University of Arizona, published in Higher Education

Summary: Based on a scientometric study of Scopus science co-publications between the U.S. and China over the past 5 years, this study found that US research article publications would have declined without co-authorship with China, whereas China’s publication rate would have risen without the USA, challenging US political rhetoric and attempts to curb international research engagement with China and demonstrating the benefits of US scientific collaboration with China for both the US nation-state and global science.

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 Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.

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

Welcome to the ChinAI Newsletter!

This week’s translation was finished a while ago, but I wanted to time it for around this time of the year, as last Monday (11/11) was Singles’ Day (光棍节) in China.

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

Feature Translation: How does AI work in the realm of love and marriage?

A Tencent Research Institute (TRI) piece, written by Lao Mu (老木, alias name) in July 2019, uses a novel initiative in Ehime Prefecture, Japan, which deploys AI to help young people find a match, to examine the possibilities of AI + dating. “AI” in the sense that it’s used throughout the piece is relatively basic matching algorithms (finding the closest match to one person on a range of different characteristics). Thanks to Riko, a good friend from Iowa City, who now works in Tokyo, for her help in adding qualifying comments to the full translation.

INTERESTING FINDINGS

• Japanese government conducted a survey of 3,980 unmarried men and women aged 20-40 in December 2018 that found 68% are “okay with not getting married,” and 60% believed "even if they are married, they are okay with not wanting children.”

• Based on the results of this survey, the Japanese government published a White Paper on “Societal Countermeasures for the Declining Birthrate” in June 2019. This White Paper noted new initiative in Ehime that claimed using AI to screen and match people, in the process of a collective blind date activity organized by the prefecture, improved the success rates of blind dates by 16%.

• The piece goes on to reflect on how to "apply “product thinking” — breaking down the complex problem of finding your match into simpler ones that can be quantified — to applying AI in the realm of love and marriage.

FULL TRANSLATION: How does AI work in the realm of love and marriage?

ChinAI Links (Four to Forward)

Should Read: Love in the Time of AI — Guardian article

A great piece that gives more context for this week’s feature translation: “When dating sims first became popular in Japan, they were often reported on by the media with a tone of moralizing disgust, partly because of the obsessive way fans played. These games were seen as an escape, a last resort for nerdy men who needed virtual girls to substitute for real, healthy heterosexual relationships. Along with anime and manga, dating sims were blamed for the low fertility rates in Japan, and the young men who played these games were sometimes described as “herbivores”, as if lacking in carnal desire. This attitude was shared by western media, too, where Japanese dating sims were seen as a curious, almost alien pathology…With the popularity of dating sims now growing outside Japan, similar concerns have once again emerged. In China, where a dating sim called Love and Producer was downloaded more than 7m times in its first month, media reports about the game have been mostly negative, if not alarmist. One Chinese commentator argued that the only reason young people were drawn to dating sims was because their real lives are “brutally lacking” in real love. “The simplicity, consumerism, and hypocrisy of romantic simulation games,” he wrote, “reflect the love-free disease that belongs to this era.”

Should-read: Hikvision Markets Uyghur Ethnicity Analytics, Now Covers Up

Charles Rollett of IPVM finds that Hikvision has marketed an AI camera that automatically identifies Uyghurs on its product page. The site was taken down after IPVM questioned them on it. Related: Recent NYT report exposes 400 pages of internal Chnese documents on the crackdown on ethnic minorities in the Xinjiang region.

Should Read: Key Takeaways from the Beijing Association of AI Conference from Hannah Kirk

Following up on last week’s readout of the BAAI conference, another attendee of the conference, Hannah Kirk, provides her summary of a panel discussing AI ethics. Seems like there was good discussion of how cultural context will shape the interpretation of AI principles that use the same words.

Should Read: National Security Commission on Artificial Intelligence Interim Report for Congress

The NSCAI released its interim report for Congress. I was impressed by the positive vision how the U.S. can leverage AI to present a positive vision of the values it stands for, especially the emphasis that ethics and strategic necessity are compatible with one another. Page 60-65 are especially impressive — based on more than 30 briefings with experts from AI-first companies, traditional companies that have successfully integrated AI, and AI orgs within the gov., NSCAI outlines a model for integrating AI across the DoD workforce.

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 Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.

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

Welcome to the ChinAI Newsletter!

We live in a crazy-wonderful world where 100+ people now pay to support the ChinAI newsletter. A good chunk of subscription fees go to supporting contributors’ translations/analysis like this week’s amazing readout of the Beijing Academy of AI Conference which took place earlier this month. As extra incentive for subscriptions: subscribers will get access to a ChinAI library document with all the ChinAI translations grouped by category (like a library!).

Jeffrey Ding@jjding99

Somehow 100+ people in this world pay to read the #ChinAI newsletter. These folks subscribe w/ no expectation of exclusive content (each #ChinAI issue is open to all). Their support goes toward a cast of contributors & my gummy worm addiction. Join them: ChinAI NewsletterJeff Ding’s weekly translations of writings from Chinese thinkers on China’s AI landscapechinai.substack.com

November 5th 2019

21 Likes

As I’ve mentioned in passing in previous issues, last year I struggled with severe neck pain. I don’t think it’s something that will ever completely go away, but I’m grateful for the days now where it doesn’t consume every waking moment, and I’m especially grateful that it made me realize that I couldn’t churn out 10,000 word translations every week anymore, and needed to rely on a community of ChinAI contributors. The payoff is a much more balanced lifestyle and the chance to feature amazing contributions/translations like this week’s.

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

Feature Translation: Beijing Academy of AI (BAAI) Conference 2019

The BAAI, established in Nov 2018 as an implementing body for the Beijing Zhiyuan Action Plan (a Beijing-specific AI plan), hosted its first global AI summit last week. A total of 1500 people attended the conference, including more than 100 top experts from the U.S., UK, Japan, Canada, Singapore, the Netherlands, and China. A ChinAI contributor who wishes to remain anonymous attended the BAAI Conference last week, and wrote up the following dispatch:

CONTEXT:

• BAAI was set up by the Beijing Municipal Science and Technology Commission and Haidian District government, with the support of the Ministry of Science and Technology and leading academic and industry players such as Peking University, Tsinghua, the Chinese Academy of Sciences, Baidu, Xiaomi and Megvii.

• More analysis on BAAI can be found in Thomas Lehmann’s piece in the recent DigiChina report “AI Policy and China.” ChinAI #52 covered BAAI’s Beijing AI ethics principles.

• The conference featured a range of technical presentations and conversations, as well as discussions on AI ethics and governance. The write-up presents excerpts from opening session speeches/slides from: Vice Minister of Science and Technology Li Meng and Vice Mayor of Beijing Yin Yong; Huang Tiejun, dean of BAAI Zhiyuan Research Institute, who gave an overview of BAAI’s first year of work; Gao Wen, Director of the Peng Cheng Lab (PCL) in Shenzhen.

MAIN TAKEAWAYS from anonymous ChinAI contributor:

• Immense efforts are being made by local governments to implement the national New Generation AI Development Plan and address shortcomings in China’s AI ecosystem. The work of BAAI and PCL is ambitious and wide-ranging, bringing together the resources and expertise of leading figures and organizations across academia and industry.

• To address shortcomings in top talent and fundamental breakthroughs, BAAI aims to provide funding (~USD 71,000 per year) to 300 AI scholars identified over three years. The dean of BAAI said in his conference speech that the youngest of the current scholars is not young enough: “The two largest contributors in the fields of computing and artificial intelligence, Turing and Gödel, made groundbreaking contributions when they were 24 years old. Therefore, we hope that more young students in the future will be able to take the lead and in the future become the backbone of China's artificial intelligence innovation.” As an aside, the UK recently announced a Turing Fellows program that similarly aims to support top AI talent.

• PCL is trying hard to reduce China’s dependency on American platforms and frameworks. PCL’s CloudBrain aims to provide shared access to an exascale supercomputer combined with an open source framework and machine learning tools. Its Zhihui code hosting site was framed in large part as a response to the risk of the US cutting off access to GitHub. However, I am sceptical that these projects will deliver significant results. As an article by Helen Toner and Lorand Laskai in a recent report by DigiChina said, “The network effects that arise because researchers want to use the same frameworks as their collaborators (and because frameworks with more users are generally better maintained over time) mean it could be increasingly difficult for a Chinese company to come from behind and dethrone established frameworks.” Exchange ‘frameworks’ for ‘code hosting platforms’ here and the same logic applies.

REFLECT FURTHER on BAAI’s and the Shenzhen-based Peng Cheng Lab (PCL) actual value-add — Jeff’s thoughts:

What is BAAI actually contributing? In his speech, the dean of BAAI’s Zhiyuan Research Institute claims that Zhiyuan Institute released the world's largest object detection data set [Objects365]. The ICCV paper that introduces this dataset is a Megvii paper (supervising author is the chief scientist at Megvii). They acknowledge support from the National Key R&D Program of China, Beijing S&T Program, and BAAI. The other two are established funding schemes, but I’m not sure what BAAI is bringing to the table here (maybe just more funding or some sort of coordinating power?). My cynical take is that it’s more of a “PR-based, let the bureaucrats believe they are doing cool AI things” contribution. BAAI claims they are planning 10 joint AI labs with industry (two already established with Megvii and Jingdong). Again what is BAAI contributing to the joint lab? Why wouldn’t Megvii and JD just partner with Tsinghua or a university instead? To be clear, I’m not saying BAAI is not doing good work on the AI ethics and safety side — we’ve covered that in a previous issue — I’m just skeptical of their value-add to leading-edge technical research.

Gao Wen, director of PCL, gives more detail on hardware deficiencies in his slides: Xilinx and Altera have nearly 90% of the FPGA market, NVIDIA have nearly 70% of the global GPU market; Boston Dynamics robot Atlas relies on the company’s huge edge in high-precision sensors and motion control algorithms. It’s not just chips, it’s also high-end sensors (e.g. analog and radio frequency sensors).

PCL is the more interesting body to me: It’s a Provincial Research Lab opened in 2018 by the Guangdong Provincial Government and funded and managed by the Shenzhen Municipal Government. In partnership with the Harbin Institute of Technology (Shenzhen), PCL also cooperates with local industry and research institutions such as the Shenzhen branches of Peking and Tsinghua Universities, the National Supercomputing Center in Shenzhen, Huawei, Tencent, and ZTE.

PCL’s CloudBrain (p. 8-10 of full translation) brings real value-add on through access to compute: CloudBrain 1 is a large-scale cluster system with 100 Petaflops of computing power, including NVIDIA GPUs, Huawei GPUs, and Cambrian AI chips. A machine of 1000 Petaflops will probably be built next year, which can be used by universities, research institutes, and SMEs for training models. The goal of 1000 Petaflows (an exaflop) is generally considered a big milestone for compute over the next few years, which the DOE is heading towards.

Still, I think the importance of these supercomputing benchmarks is often overstated. It often depends on whether the “supercomputer” is defined as a fully integrated supercomputer with low latency or a bunch of distributed compute. CloudBrain seems to fit the latter case: at present, the distributed computing resources connected to CloudBrain 1 include Zhongshan University’s Tianhe-2 supercomputer in Guangzhou, Hefei Brain-Inspired Computing Center’s server cluster (which they spent nearly 100 million RMB on), and PCL’s supercomputing center in Shenzhen.

***The full translation contains the anonymous ChinAI contributor’s notes from the conference, extracts from speeches, and translations of presentation slides:

FULL TRANSLATION: BAAI 2019 Conference Write-Up

ChinAI Links (Four to Forward)

Must-read: The Early History of AI in China (1950s-1980s) by Jieshu Wang

Jieshu, PhD student in the Human and Social Dimensions of Science and Technology Program at ASU, takes us through the tumultuous development of cybernetics and AI in China, including a fascinating section on how Qian Xuesen (father of China’s space program) was convinced that AI was associated with “exceptional human body functions (e.g. ESP, telepathy). Jieshu argues that the establishment of the Chinese Association for Artificial Intelligence in 1981, the first official research institute in China dedicated specifically to AI, was critical to the development of AI in China. The piece is a welcome reminder for us to not get caught up in the presentist view of AI in China.

Should Read: Scary China by Yuan Yang in Chinese Storytellers Newsletter

Reflecting on the development of the “Scary China” narrative, Yuan (FT’s China tech correspondent) writes “We as journalists should take some responsibility for the rise of this meme. As an example, I recall that it was early 2017 when the phrase “AI arms race” started to gain prominence. Although I can’t remember exactly how the phrase entered the collective consciousness of our English-speaking newsrooms, I can guess at the reason we all reached for it: it sounded scary. Scary things are important, and important things are newsworthy.” The question she poses to the Chinese Storytellers group: How do we create narratives that extend our emotional palettes beyond fear?

Should Read: AI Principles: Recommendations on the Ethical Use of Artificial Intelligence by the Department of Defense

Longtime readers of ChinAI know that I’m not the biggest fan of what I lovingly call the military industrial complex. But this Defense Innovation Board (DIB) document on AI principles is worthy of praise, drawing on a 15-month study that involved public listening sessions at major universities. The 74-page supplementary document draws attention to emergent effects (unintended escalation and speed) of AI systems, which are especially important in my opinion. Full disclosure: I participated in a stakeholder interview session, providing input to the DIB on this initiative.

Should Read: Assessing the State of AI R&D in the US, China, and Europe — Part 1: Output Indicators

Stefan Torges provides a clear-eyed comparison between the state of AI R7D in the U.S. and China, finding that for almost all indicators the US institutions are the global leaders. I particularly liked the comparisons between the assumptions and methodologies driving various indicators. Stefan also draws from my testimony before the U.S.-China Economic and Security Review Commission, which provides a systematic framework to assessing the fuzzy concept of “national AI capabilities.”

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 Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.

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

Welcome to the ChinAI Newsletter!

Our subscription drive continues: we’re so close to 100 subscribers — if we hit that mark, I’ll give subscribers access to an organizational document with the complete collection of ChinAI translations to date (grouped by theme/category).

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

Feature Translation: Qianzhan Chanye Report on China’s AI Industry

CONTEXT: Based on what I’ve read, Qianzhan’s (a consultancy/research institute) 50-page slide deck is the best open-source, overview-style report on China’s AI industry. I’d wager it’s probably better than the closed-source reports from Goldman, government departments, etc. I’m going to keep harping on this point: If you’re following China’s tech sector, you will miss a lot if you’re not looking at the output of orgs like Qianzhan, CCID, and other Chinese research firms, as these places are only going to scale up and get better.

• I only had time this week to do a quick scan of the first 18 slides and did a few quick-and-dirty translations of interesting sections (if there’s no comment on the slide that means I probably didn’t do any translations on it).

• ****I’ll circle back to this every now and then to translate more chunks, but if you’re keen to help out with the whole thing, let me know! In a previous issue, an anonymous ChinAI contributor and I translated a 5G PPT slide deck that went viral in China and became recommended reading for all Huawei staff. There’s been lots of good discussion happening in the comments of that translation, and I hope the same happens for this one.

THE ESSENTIALS:

• Three-layer (Foundation-Technology-Application) division of China’s AI industry : Slide 5’s specification of the different layers of AI is the first clear sign that this is good stuff. Representative companies in the foundation layer (slide 6) includes usual suspects like Cambricon, Horizon Robotics in chips; Baidu, Haiyun Shuju in big data; BAT in cloud computing, Huawei in 5G.

• What stood out about the representative companies in various layers of China’s AI industry (slide 6-8): Datatang is listed as a key big data company. We analyzed their involvement in a turning point case in personal information protection in ChinAI #19. Ultrapower shows up in more verticals than Alibaba and Tencent in the foundation layer — we did a ChinAI Company Profile of them in ChinAI #63.

• Ping An is the most important company in China’s AI landscape that you’ve never heard of: nominally an insurance company but transforming quickly into a tech company, they are everywhere in the foundation (including cloud computing) and application layers. However, they don’t show up in any of the verticals in the technology layer.

• Slide 18 on China’s AI talent Training System was particularly instructive:

It’s a nice overview of industry-academia joint AI institutes. We covered the iFlytek-Chongqing University of Posts and Telecommunications one in ChinAI #7. These types of schools may be just as important as the Tsinghuas and Pekings for the diffusion of AI advances across the whole country.

TRANSLATION IN PROGRESS: Qianzhan Chanye Report on China’s AI Industry

ChinAI Links (Four to Forward)

Must-read: DigiChina Special Report: AI Policy and China — Realities of State-Led Development (Edited by Graham Webster)

Big ups to Graham Webster and the Stanford-New America DigiChina team for an excellent report on China’s AI Policy. Each one of the pieces are worth a read, and I was impressed by the consistent effort to translate and analyze Chinese-language sources throughout. I learned a lot in particular from Siodhbhra Parkin’s piece on how AI can better serve people with disabilities in China (p. 33 of the report).

Must-listen: ChinAI Pod #2: Reframing Superintelligence with Eric Drexler

Nick Bostrom’s book Superintelligence is what got me really interested in AI in the first place; after joining FHI a couple years later, Eric Drexler’s report Reframing Superintelligence significantly shifted my view of superintelligence in a way that lived up to the report’s title. It was an honor to talk with the founding father of nanotechnology about his framework of Comprehensive AI Services.

If you like what you hear, pass it on! If you don’t like what you hear but are willing to take a chance on hearing more, let me know in a public setting so that I get the feedback and also promo for the pod. We should be on all the places you get your podcasts now (just search “ChinAI,” but if we’re not, you can paste this link into your podcast app and the feed should pop up: https://api.substack.com/feed/podcast/2660/private/df055a23-b213-4016-9d0b-cbd168059b40.rss

Should-read: Translations related to China’s AI Policy by Center for Security and Emerging Technology

Ben Murphy, Chinese STEM Translation Lead at CSET, is doing fine work with these full text translations. A few notes from some of the translations that CSET has published. I plan to circle back every month or so and highlight what’s interesting from the translations.

1.Translation of Project to Strengthen Development of the Defense Technology Industry at the Grassroots Level: Guidelines for Basic Research and Cutting-Edge Technology Projects (2018) — a notice to Chinese universities and research institutes on emerging technologies (including AI-related domains) that the PLA prioritizes, from SASTIND (a civilian agency that funds research in support of PLA requirements)

• AI features heavily in three of the six themes: intelligent detection/identification and autonomous control technology; brain-machine intelligence and biological interdisciplinary technologies; highly reliable information security and new types of technology

• Nothing in first 2 themes was that surprising (cloud-based intelligent target recognition and tracking, intelligent decision-making for virtual battlefield environments, intelligentized exoskeleton technology); the third theme was more interesting to me and included key tasks such as “intelligent mining technology” for firmware vulnerabilities as well as big data analysis-based intelligent fixed decryption technology for electronic documents

2.Translation of Tianjin Municipal Action Plan for Military-Civil Fusion Special Projects in Intelligent Technology (2018) — Wow this is a rich document with a lot of material in the translator’s footnotes

• There’s a story/report waiting to be written about Tianjin’s technology policy (remember: Tianjin was one of the first cities to set up a big AI fund. “While many Chinese local governments have published military-civil fusion plans, Tianjin’s is among the most detailed,” per CSET’s summary.

• A lot of the initiatives in here sound very ambitious (“build an authoritative IoT perception testing and authentication center for the state, military, and industrial sectors in Tianjin’s Binhai New District!”), but I’m also reminded of the empty promise of the Tianjin Eco-city.

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 Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.

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