A research team from the University of Southern California and Google proposes TOME, a “mention memory” approach to factual knowledge extraction for NLU tasks. A transformer model with attention over a semi-parametric representation of the entire Wikipedia text corpus, TOME can extract information without supervision and achieves strong performance on multiple open-domain question answering benchmarks.
Nvidia’s paper Large Scale Language Modeling: Converging on 40GB of Text in Four Hours introduces a model that uses mixed precision arithmetic and a 32k batch size distributed across 128 Nvidia Tesla V100 GPUs to improve scalability and transfer in Recurrent Neural Networks (RNNs) for Natural Language tasks.
Benjamin Sanchez-Lengeling from Harvard University and Alán Aspuru-Guzik from the University of Toronto have successfully applied machine learning models to speed up the materials discovery process. Their paper Inverse molecular design using machine learning: Generative models for matter engineering was published July 27 in Science Vol. 361.
The Internet is woven into our everyday lives. We access massive amounts of data through our laptops, smartphones and tablets. This free flow of information however has prompted attempts to filter content which may not be appropriate for example for young people. One such new effort from Brazil puts virtual bikinis on nudes.
Shanghai-based Tianrang Intelligence announced today that it has secured over US$26 million in Series A funding led by CMB International Capital and Gaorong Capital, with Ziniu Fund and Wacai Technology also participating. The two-year-old startup is working on general artificial intelligence solutions and will use the new funding to create a “City Brain Lab.”
The Technical University of Munich (TUM) had lured Cheng from a Japanese research institute. At TUM he founded the Institute of Cognitive Systems (ICS). With eight employees in a central office on Karlstraße 45, Cheng set to work on his arduous task: recreating the complexities of human skin and wiring it all to a brain.
Neural networks can be notoriously difficult to debug, but a Google Brain research team believes it may have come up with a novel solution. A paper by Augustus Odena and Ian Goodfellow introduces Coverage-Guided Fuzzing (CGF) methods for neural networks. The team also announced an open source software library for CGF, TensorFuzz 1.
Guangzhou-based Xiaopeng Motors (Xpeng) announced today that it had scored CN¥4 billion (US$585 million) in its Series B+ funding round. This is the largest 2018 single-round funding total among Chinese Electric Vehicle startups (NIO, Byton, WM Motor, etc.), and pushes Xpeng’s valuation to CN¥25 billion (US$3.66 billion).
Since 2010, the annual ImageNet Large-Scale Visual Recognition Challenge has been the most widely recognized benchmark for testing image recognition algorithms. Tencent Machine Learning picks up the challenge with its new paper Highly Scalable Deep Learning Training System with Mixed-Precision: Training ImageNet in Four Minutes.
Scientists at the National Supercomputing Center of Tianjin in China have unveiled their prototype of a next-generation exascale supercomputer. The centre says the Tianhe-3 (天河3号) can perform at one exaFLOPS, or a billion billion calculations per second— the threshold for coveted “exascale” status.
Over the past year Chinese startups and tech giants alike have developed their own voice technology ecosystems comprising product manufacturers, tech solution providers, content makers, and platforms designed to help hardware or skill applications “speak and be smarter.” In this article, we look at the current state of open voice platforms in China.
At the annual Google Cloud Next conference which kicked off July 24 in San Francisco the company unveiled a series of AI-based product releases and enhancements for its analytics and machine learning tools, additional applications on G Suite, and new IoT products.
Meituan, one of China’s biggest group buying service providers, today announced an ambitious strategy to support autonomous delivery vehicles: “The company is going to use self-driving technologies to build the world’s largest interactive platform for intelligent delivery, and to speed up the development of unmanned delivery technology.”
Chinese IoT startup Tuya Smart has completed its Series C funding with nearly US$200 million, led by Future Fund, NEA, China Broadband Capital, Quadrille Capital and C.M.Capital. The company will apply the funds to its AI+IoT, Smart City, Smart Business and Smart Home technologies.
The dearth of AI talents capable of manually designing neural architecture such as AlexNet and ResNet has spurred research in automatic architecture design. Google’s Cloud AutoML is an example of a system that enables developers with limited machine learning expertise to train high quality models. The trade-off, however, is AutoML’s high computational costs.
Facebook is working with a select group of advertisers to create augmented reality ads for its News Feed. When users activate an ad’s tap-to-try AR capability it can display for example how a pair of glasses would look on their face via the user’s webcam and screen. Facebook says it also intends to expand shopping support in Instagram Stories.
Four years after its founding, human-computer interaction and robotics company Rokid held its first major press conference on June 26 in Hangzhou. Dubbed “Rokid Jungle,” Founder and CEO Misa Zhu quipped that the presser was “a combination of Google IO and Apple WWDC.”
Last week, Singh posted a YouTube video, Making Amazon Alexa respond to Sign Language using AI, in which he smoothly communicates with Alexa using sign language. Alexa, which is installed in a laptop with a built-in camera, interprets Singh’s signed queries in real-time, converting them into text and delivering appropriate responses.
On July 10th, with German Chancellor Angela Merkel and Chinese Premier Li Keqiang looking on, Siemens AG signed a partnership agreement with Alibaba Cloud — the cloud-computing arm of Internet conglomerate Alibaba — to bring an Industrial Internet of Things (IIoT) upgrade to China’s manufacturing industry.
Xilinx, the world’s leading designer and supplier of programmable logic devices, today announced its acquisition of DeePhi Tech — a Beijing-based chip startup with a focus on machine learning, specializing in deep compression, pruning, and system-level optimization for neural networks.
The country is aiming to build the world’s top hub for AI innovations and talents, and yesterday, prestigious Tsinghua University issued a progress report. The 2018 China Artificial Intelligence Development Report provides an insightful, data-intensive overview of the current status of AI in China.
From Hayao Miyazaki’s Spirited Away to Satoshi Kon’s Paprika, Japanese anime has made it okay for adults everywhere to enjoy cartoons again. Now, a team of Tsinghua University and Cardiff University researchers have introduced CartoonGAN — an AI-powered technology that simulates the styles of Japanese anime maestri from snapshots of real world scenery.
Innoviz is a two-year-old Israeli provider of LiDAR remote sensing solutions. The company took a leap forward in the solid state LiDAR field recently by scoring a couple of important partnerships: BMW chose solid-state Innoviz LiDAR sensors and computer vision tech to enable its Level 3-5 autonomous vehicles in April…
The 41st International ACM SIGIR Conference on Research and Development in Information Retrieval is now in full swing at the University of Michigan campus in Ann Arbor, US — although some leading researchers won’t be attending. Yesterday the respected conference presented its Best Paper Awards.
When robots need a brain, their creators turn to Silicon Valley. Most of the AI tech that drives advanced robotics originates in Bay Area labs. At last month’s ReWork Deep Learning for Robotics Summit in San Francisco, researchers from Silicon Valley AI labs and institutes discussed their latest work and how it is being used to teach robots. Synced was onsite to bring you an inside look at their work.