• Chinese Tradition Inspires Machine Learning Advancements, Product Contributions

    Posted by George Thomas Jr.

    Machine learning solves Chinese riddles

    A new online Chinese riddle game is steeped in more than just tradition. In fact, the machine learning and artificial intelligence that fuels it derives from years of research that also helps drive Bing Search, Bing Translator, the Translator App for Windows Phone, and other products.

    Launched in celebration of the Chinese New Year, Microsoft Chinese Character Riddle is based on the two-player game unique to Chinese traditional culture and part of the Chinese Lantern Festival. Developed by the Natural Language Computing Group in the Microsoft Research's Beijing lab, it resulted from a challenge between colleagues to build a computer that could solve riddles, and, in turn, has led to numerous machine translation advancements and patents.

  • Microsoft Research Expands Microsoft Band Productivity Functionality

    Posted by Microsoft Research

    Band gets machine-learning enhancements

    When Microsoft Band debuted in October 2014, it was more than a wearable fitness device. In addition to valuable health-related features such as a 24-hour heart-rate monitor, Guided Workouts, and built-in GPS tracking, it allowed users to preview their emails and texts, get calendar alerts, get actionable insights from Microsoft Health, and connect to the Cortana digital assistant.

    Users caught on quickly—and immediately clamored for more productivity features, including the ability to respond quickly and naturally to urgent messages. Thanks to cutting-edge technologies incubated by Microsoft researchers, a variety of new features have been introduced for Windows Phone 8.1 users in the latest update to Microsoft Band, which was released on Monday. They include an innovative virtual keyboard designed specifically for the device and new voice reply functionality powered by Cortana.

  • Machine Learning Gets Big Boost from Ultra-Efficient Convolutional Neural Network Accelerator

    Posted by Doug Burger

    Doug Burger

    (Editor’s Note: Doug Burger, a processor architect by training, is a Microsoft researcher focused on disrupting the very fabric of datacenter processing power in a mobile-first, cloud-first world.)

    I’m excited to highlight a breakthrough in high-performance machine learning from Microsoft researchers.

    Before describing our results, some background may be helpful. The high-level architecture of datacenter servers has been generally stable for many years, based on some combination of CPUs, DRAM, Ethernet, and disks (with solid-state drives a more recent addition). While the capacities and speeds of the components—and the datacenter scale--have grown, the basic server architecture has evolved slowly. This slow evolution is likely to change, however, as the decelerating gains from silicon scaling are opening the door to more radical changes in datacenter architecture.

  • 2015 Oscars Prediction Model Navigates Backlash, Outrage, Shifting Sentiment

    Posted by George Thomas Jr.

    Ava DuVernay

    While it may be difficult to name a year in which The Oscars were not embroiled in some controversy regarding nominees, this year seems particularly pronounced, with widespread criticism spanning numerous categories, mainly objecting to the lack of nominations for “Selma,” and, on a light-hearted note, incredulity about “The Lego Movie” not receiving an Animated Feature Film nomination.

    To mis-paraphrase a song from that excluded film, everything is not awesome in Oscar-nomination land.

    Yet, as Microsoft researcher David Rothschild (@DavMicRot) notes, that doesn’t mean the un-nominated can’t affect the Oscar winners in those “disputed” categories. “Absence in the category makes a difference in the distribution of votes for the remaining choices,” he says.

  • Microsoft Researchers' Algorithm Sets ImageNet Challenge Milestone

    Posted by Richard Eckel

    Jian Sun

    The race among computer scientists to build the world’s most accurate computer vision system is more of a marathon than a sprint.

    The race’s new leader is a team of Microsoft researchers in Beijing, which this week published a paper in which they noted their computer vision system based on deep convolutional neural networks (CNNs) had for the first time eclipsed the abilities of people to classify objects defined in the ImageNet 1000 challenge.