The U.S. may be trailing behind China in artificial intelligence (AI) research -- or at least in journal articles that mention "deep learning" or "deep neural network" -- according to the White House's National Artificial Intelligence Research and Development Strategic Plan. While the U.S. remains an early leader in deep learning research, China seems to be spending more time studying the technology and making influential contributions to the field than we are, according to White House's research, leading the U.S. in both the number of published deep learning studies and the number of studies cited by other researchers. The administration's plan proposes future research and development projects in the field of AI and was released in conjunction with Preparing for the Future of Artificial Intelligence, a report that overviews the current and future state of AI technology and its place in society. Both papers were shared in anticipation of the White House Frontiers Conference. Both countries, obviously, are devoting a lot of attention to AI and deep learning, in particular.

Word embedding is a technique that treats words as vectors whose relative similarities correlate with semantic similarity. This technique is one of the most successful applications of unsupervised learning. Natural language processing (NLP) systems traditionally encode words as strings, which are arbitrary and provide no useful information to the system regarding the relationships that may exist between different words. Word embedding is an alternative technique in NLP, whereby words or phrases from the vocabulary are mapped to vectors of real numbers in a low-dimensional space relative to the vocabulary size, and the similarities between the vectors correlate with the words' semantic similarity. For example, let's take the words woman, man, queen, and king.

As restrictions prevent foreign internet companies like Google and Facebook from operating in China, Chinese technology companies have free access to its users, resulting in access to more data, which in AI research means better results. China is also considering military applications for AI, most notably the development of next-generation "fire-and-forget" cruise missiles equipped with AI. Not to be left behind in the AI race, the White House established the NTSC Subcommittee on Machine Learning and Artificial Intelligence in May this year, which in turn created the National Artificial Intelligence Research and Development Strategic Plan. This plan advocates creating a better setting in the US to enable growth in AI research. Increased support for AI research in the US will aid key players like Amazon, Microsoft, Google, Apple, and IBM.

While many advances have been made, it still has been difficult for non-programmers to create objects made of many materials (or mixtures of materials) without a more user-friendly interface. But this week, a team from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) will present "Foundry," a system for custom-designing a variety of 3-D printed objects with multiple materials. "In traditional manufacturing, objects made of different materials are manufactured via separate processes and then assembled with an adhesive or another binding process," says PhD student Kiril Vidim?e, who is first author on the paper. "Even existing multi-material 3-D printers have a similar workflow: parts are designed in traditional CAD [computer-aided-design] systems one at a time and then the print software allows the user to assign a single material to each part." In contrast, Foundry allows users to vary the material properties at a very fine resolution that hasn't been possible before.

We propose a unified and systematic framework for performing online nonnegative matrix factorization in the presence of outliers. Our framework is particularly suited to large-scale data. We propose two solvers based on projected gradient descent and the alternating direction method of multipliers. We prove that the sequence of objective values converges almost surely by appealing to the quasi-martingale convergence theorem. We also show the sequence of learned dictionaries converges to the set of stationary points of the expected loss function almost surely. In addition, we extend our basic problem formulation to various settings with different constraints and regularizers. We also adapt the solvers and analyses to each setting. We perform extensive experiments on both synthetic and real datasets. These experiments demonstrate the computational efficiency and efficacy of our algorithms on tasks such as (parts-based) basis learning, image denoising, shadow removal and foreground-background separation.

Years of evolution got something right. Scientists at the University of Tokyo found that, in trying to keep their humanoid robot cool, the most effective way to avoid overheating was to make it sweat. Kengoro is made up of 108 motors, with a frame laser sintered from aluminum, and researchers found that making the robot sweat was an effective cooling system for a bot filled to the brim with bolts. Even with the space-saving technique, the bot weighs 123 pounds and stands 5 feet 7 inches tall. "Usually the frame of a robot is only used to support forces," lead author Toyotaka Kozuki told IEEE Spectrum in an interview published Friday.

Alibaba will be among 13 businesses working with the Hangzhou government on a'brain' for the city and will work with the National Astronomical Observatory of China (NAOC) on deep space exploration projects, it announced at its annual Computing Conference this week. According to the retail and cloud computing giant, it will be supplying a range of its tech services such as AI, deep learning and data storage. The B2B technology supply side to the Alibaba business is growing fast and puts it very much in battle with Amazon on a global playing field. The Hangzhou City Brain project is a new government initiative to address its urban city living issues, such as traffic congestion. It will use Alibaba Cloud's AI program "ET" and big data analytics capabilities to perform real-time traffic prediction by using its video and image recognition technologies.

Just about every service industry--from retailers to restaurants to hotels--has developed some kind of robot to attend to your needs on the cheap. The latest effort in the banking world (there are already robotic bank receptionists in China and Japan) is to take the rote responses of bots to the next level, by adding a touch of human empathy. The Royal Bank of Scotland (paywall) plans to unveil its new artificial intelligence system, known as "Luvo," by the end of the year. The AI service, designed by IBM, will attend to customer banking needs through its mobile or online as a chatbot. It will function similarly to Siri, the iPhone virtual assistant that answers questions with a distinct voice and "personality."

Tencent, the Chinese tech giant that owns League of Legends, Supercell Games and WeChat, the most popular messaging app in China, announced on Friday that it will release a consumer quadcopter by the end of October. The Ying drone will cost 300 and offer the option to capture 4K video with the onboard storage or livestream it through a WeChat conversation in 720p resolution. It's tiny, weighing less than a pound and able to fit easily in a backpack. Tencent reportedly teamed up with Qualcomm and Chinese drone manufacturer Zerotech to produce the Ying. Of course, DJI's new offering costs more than three times as much, so there likely won't be a whole lot of market overlap between the models.

Huawei will provide a US 1 million fund to UC Berkeley for research into many subjects of interest in AI, including deep learning, reinforcement learning, machine learning, natural language processing and computer vision. Through close cooperation, the Research and Development (R&D) teams of Huawei and the Berkeley Artificial Intelligence Research (BAIR) Lab will strive to make significant breakthroughs in AI theories and key technologies. The two parties believe that this strategic partnership will fuel the advancement of AI technology and create completely new experiences for people, thus contributing greatly to society at large. As one of the world's leading higher education institutes, UC Berkeley has profound expertise in machine learning and other AI domains. Its newly founded BAIR Lab brings together UC Berkeley researchers across the areas of computer vision, machine learning, natural language processing, robotics, and research planning.