China’s AI research quickly catching up to US in this new metric: number of citations

… for 10 of the top 11 AI unicorns – private companies with a valuation of US$1 billion or above – according to a separate compilation by CB Insights.

China’s rapid ascendance in artificial intelligence research continues, supported by new data that suggests the country is catching up to the US in high impact research in the field.

While China has already surpassed the US in number of published AI papers, the country’s AI researchers are poised to be in the top 50 per cent of most cited papers this year and in the top 10 per cent next year, according to findings by the Allen Institute for Artificial Intelligence, a Seattle-based non-profit that conducts research and engineering projects.

The results, which drew on a database of 2 million papers published up to 2018, show that the US share of citations in the top 10 per cent of AI papers has declined gradually from 47 per cent in 1982 to 29 per cent last year. China, on the other hand, has risen to over 26 per cent of citations in 2018.

“Citation counts are a lagging indicator of impact, so our results may understate the rising impact of AI research originating in China,” said the report that was published on Wednesday.

The findings highlight the latest efforts made by the world’s two superpowers in what has been dubbed the fourth industrial revolution, and provides a new gauge for academics to measure strength in AI research.

Xiaomi steps up AI emphasis in new reshuffle

Last month, US President Donald Trump issued an executive order on maintaining “American leadership in AI”. The initiative, which directed federal agencies to prioritise investments in research and development in the field, was seen by some as a response to China’s ambition to create a domestic AI industry worth 1 trillion yuan (US$147 billion) and become a global AI powerhouse by 2030.

The world’s second largest economy is on a mission to roll out AI in all walks of life, from catching jaywalkers and saving toilet paper by using facial recognition to more loftier applications such as self-driving cars and medical diagnosis.

China has led the world in the number of patent filings in AI since 2014, followed by the US, according to the World Intellectual Property organisation. The two countries are home to the majority of heavyweight AI start-ups, accounting for 10 of the top 11 AI unicorns – private companies with a valuation of US$1 billion or above – according to a separate compilation by CB Insights.

China overtook the US in the number of AI research papers in 2006, well before Beijing unveiled its 2030 AI blueprint in 2017, according to the Allen Institute for Artificial Intelligence.

As Chinese researchers are sometimes “stereotyped as making incremental research contributions”, the institute said its data provides an alternative to ranking by sheer volume of AI papers published.

“By projecting current trends, we see that China is likely to have more top 10 per cent papers by 2020 and more top 1 per cent papers by 2025,” it said.

More from South China Morning Post:

This article China’s AI research quickly catching up to US in this new metric: number of citations first appeared on South China Morning Post

For the latest news from the South China Morning Post download our mobile app. Copyright 2019.

Related Posts:

  • No Related Posts

Towards Literate Artificial Intelligence

Over the past decade, the field of artificial intelligence (AI) has seen striking developments. Yet, today’s AI systems sorely lack the essence of human …

Towards Literate Artificial Intelligence

par

Mrinmaya Sachan

Ph.D. candidate in the Machine Learning Department in the School of Computer Science at Carnegie Mellon University

Vendredi 15 mars, 10:30-12:00, Salle 3195, Pavillon André-Aisenstadt

Université de Montréal, 2920 Chemin de la Tour

Résumé:

Over the past decade, the field of artificial intelligence (AI) has seen striking developments. Yet, today’s AI systems sorely lack the essence of human intelligence i.e. our ability to (a) understand language and grasp its meaning, (b) assimilate common-sense background knowledge of the world, and (c) draw inferences and perform reasoning. Before we even begin to build AI systems that possess the aforementioned human abilities, we must ask an even more fundamental question: How would we even evaluate an AI system on the aforementioned abilities? In this talk, I will argue that we can evaluate AI systems in the same way as we evaluate our children – by giving them standardized tests. Standardized tests are administered to students to measure the knowledge and skills gained by them. Thus, it is natural to use these tests to measure the intelligence of our AI systems. Then, I will describe Parsing to Programs (P2P), a framework that combines ideas from semantic parsing and probabilistic programming for situated question answering. We used P2P to build systems that can solve pre-university level Euclidean geometry and Newtonian physics examinations. P2P achieves a performance at least as well as the average student on questions from textbooks, geometry questions from previous SAT exams, and mechanics questions from Advanced Placement (AP) exams. I will conclude by describing implications of this research and some ideas for future work.

Biographie :

Mrinmaya Sachan is a Ph.D. candidate in the Machine Learning Department in the School of Computer Science at Carnegie Mellon University. His research is in the interface of machine learning, natural language processing, knowledge discovery and reasoning. He received an outstanding paper award at ACL 2015, multiple fellowships (IBM fellowship, Siebel scholarship and CMU CMLH fellowship) and was a finalist for the Facebook fellowship. Before graduate school, he graduated with a B.Tech. in Computer Science and Engineering from IIT Kanpur with an Academic Excellence Award.

Website: sites.google.com/site/mrinsachan.

Related Posts:

  • No Related Posts

Artificial intelligence progress gets gummed up in silos and cultural issues

Silos have always been considered a bad thing for enterprise IT environments, and today’s push for artificial intelligence and other cognitive …

Silos have always been considered a bad thing for enterprise IT environments, and today’s push for artificial intelligence and other cognitive technologies is no exception. A recent survey shows fewer than 50% of enterprises have deployed any of the “intelligent automation technologies” — such as artificial intelligence (AI) and robotic process automation (RPA). IT leaders participating in the survey say data and applications within their companies are too siloed to make it work.

clouds-over-chicago-cropped2-nov-2015-photo-by-joe-mckendrick.jpg

Photo: Joe McKendrick

That’s the gist of a survey of 500 IT executives, conducted by IDG in partnership with Appian. The majority of executives, 86%, say they seek to achieve high levels of integration between human work, AI, and RPA over the coming year. The problem is they have a long way to go — at this time, only 12% said their companies do this really well today.

Where are the problems? Two-thirds of executives, 66%, stated that they “have difficulty integrating existing IT investments and skills with demanding AI and RPA technology.” Notably, 43% cite changing the IT culture as an obstacle to AI and RPA. While the survey report’s authors did not spell out what kind of changes were required, it can be assumed that IT culture is hampered by a need for a constant maintenance and firefighting, versus focusing more on innovation. There may be also issues with communication between the IT and business sides of the house — as well as interacting more with data science types. Some of these issues may eventually see some relief through agile and DevOps initiatives.

Additional issues that hold back AI and RPA progress include concerns about security, cited by 41%, and application development issues seen by 34% of the group. Again, this was not elaborated in the report, but application development roadblocks likely stem from lack of proper tools to build AI-driven applications, along with the need for skills development or refreshes.

In addition, linking automation efforts to improving customer experiences was problematic. Two-thirds of executives, 66%, say they Needed a Better Multi-Channel Buying Experience. However, 26% lack the systems to deliver integrated multi-channel customer experiences, and 22% need to build or buy software to implement multi-channel customer experiences. Another 21% say they even lack a strategy for delivering integrated multi-channel customer experiences.

At this point, it appears AI and RPA are mainly the tools of the largest corporations with humongous IT staffs. While there are deployments of individual emerging automation technologies, a lack of strategy and clear alignment to business goals is resulting in siloed deployments and overwhelmed internal application development teams. Less than half of surveyed companies have deployed any form of intelligent automation. Fully half of those companies boast IT staffs in excess of 20,000 employees.

Related Posts:

  • No Related Posts

SkoltechZaryadye: Can artificial intelligence replace doctors?

Vinod Khosla, Silicon Valley’s major venture investor, believes that by 2035, all human doctors will be replaced by robots. Can he be right? There is …

Vinod Khosla, Silicon Valley’s major venture investor, believes that by 2035, all human doctors will be replaced by robots. Can he be right? There is every indication that he can: today, machine learning systems are producing stunning results in image recognition and processing techniques which are widely used in X-ray, MRI, ECG and ultrasound tests that amount to 90% of all the medical data. AI-based solutions trained on thousands of medical images can diagnose 50 eye diseases with nearly 100% accuracy and outperform qualified doctors in detecting diabetes based on a retinal image.

4000x2250_2603

Join the lecture by Skoltech Professor Mikhail Belyaev to find out where robots can fully replace humans or operate as an assistant handling routine tasks, such as initial data processing and analysis. The lecturer will share some success stories displaying the synergy of human and artificial intelligence in diagnosing various diseases, including brain pathology.

Mikhail Belyaev, Professor at the Skolkovo Institute of Science and Technology and an AI and ML expert with a ten-year track record. Prior to joining Skoltech, Belyaev worked at DATADVANCE where he was involved in the development of customized data analysis algorithms for Airbus, Toyota, Areva and other companies. He has been specializing in the analysis of medical data and primarily, data on human brain-related pathologies, since 2015.

Suitable for 16+

The lecture will be given in Russian at Zaryadye Park Lecture Center on March 26, 2019 at 7 p.m.

Registration: https://www.zaryadyepark.ru/schedule/62103/

Contact information:

Skoltech Communications

+7 (495) 280 14 81

*protected email*

Tweet about this on Twitter0Share on Facebook0Pin on Pinterest0Share on Tumblr0Share on VK*protected email*

Related Posts:

  • No Related Posts

CYBERCOM Seeks Troops Who Can Unleash Artificial Intelligence

The Defense Department’s cyber warriors shouldn’t be too concerned about artificial intelligence taking their jobs, according to their commander.

The Defense Department’s cyber warriors shouldn’t be too concerned about artificial intelligence taking their jobs, according to their commander. Instead, U.S. Cyber Command is looking for troops able to wield AI like a weapon.

During a budget hearing Wednesday held by the House Armed Services Subcommittee on Intelligence and Emerging Threats and Capabilities, Rep. Anthony Brown, R-Md., asked the Pentagon’s cyber leadership whether AI could help reduce the demand for cyber talent.

“AI and machine learning certainly has a place as we look at some of the activities that we are doing day in and day out,” CYBERCOM Commander Gen. Paul Nakasone told the subcommittee. “But I would offer, the people that make AI go, the people who make sure that our algorithms are right for machine learning, they’re the folks that I’m most focused on.”

Nakasone referred to what he called the “10x or 20x folks” that can utilize these advanced systems as a multiplier for the work being done by the command’s 133 Cyber Mission Force teams, which includes “amplify[ing] military lethality and effectiveness,” according to the Pentagon’s Defend Forward strategy released in September.

In that sense, AI won’t be replacing cyber troops one to one. But that force multiplier can ease the burden when there are gaps in the ranks.

The Cyber Excepted Service, or CES, system—established by Congress in 2016 to speed recruitment and retention by offering added incentives such as higher pay scales—has been another useful tool, particularly in speeding the hiring process, Nakasone said.

Since the first phase began in August 2017, CYBERCOM has seen a 60 percent drop in the hiring timeline, from 111 days to 44, Nakasone said, citing internal metrics.

The general said the military services have been an important asset in helping CYBERCOM recruit those people. But retaining quality cyber warriors who can make far more money in the private sector is the challenge “that’s most impactful for us,” he said.

Related Posts:

  • No Related Posts