Der deutsche Education Blog

Microsoft Research Connections Blog

The Microsoft Research Connections blog shares stories of collaborations with computer scientists at academic and scientific institutions to advance technical innovations in computing, as well as related events, scholarships, and fellowships.

  • Microsoft Research Connections Blog

    How might climate change affect our food supply?


    How might climate change affect our food supply?It’s no easy question to answer, but prudence demands that we try. Thus, Microsoft and the United States Department of Agriculture (USDA) have teamed up to tackle “food resilience,” one of several themes that make up the White House’s Climate Data Initiative.

    “Through his Climate Data Initiative, President Obama is calling for all hands on deck to unleash data and technology in ways that will make businesses and communities more resilient to climate change,” said John P. Holdren, President Obama’s Science Advisor. “The commitments being announced today answer that call by empowering the U.S. and global agricultural sectors with the tools and information needed to keep food systems strong and secure in a changing climate.”

    The Climate Data Initiative has unleashed a torrent of climate-related data from NOAA, NASA, the US Geological Survey, US Department of Defense, and other federal agencies, including the USDA. These facts and figures, which reside on's Climate website, pose a classic “big data” challenge: how to efficiently analyze enormous information sets and share the meaningful insights.

    The overarching goal is to discover the food supply's key vulnerabilities and inherent resiliency to climate change.
    The overarching goal is to discover the food supply's key vulnerabilities and inherent resiliency to climate change.

    Microsoft has posted the USDA datasets to the Microsoft Azure Marketplace (enter search term USDA), and, together with the USDA, we will be sponsoring workshops, webinars, and “appathons” to demonstrate the value of open access data and to promote the development of tools for understanding these datasets. The overarching goal is to encourage data providers, scientists, farmers, food producers and the public to discover the food supply’s key vulnerabilities and inherent resiliency. This predictive information will inform a planning model built on the powerful business intelligence tools that are part of the Microsoft Azure cloud-computing platform, enabling federal agencies, along with the public, access and tools to promote data synthesis with other data sources.

    To advance this effort even further, Microsoft Research is announcing a special Climate Data RFP focused on food resilience in the face of climate change. This RFP offers 12 months of free cloud-computing resources to 20 awardees selected from proposals submitted by September 15, 2014. Each award provides up to 180,000 hours of cloud-computing time and 20 terabytes of cloud storage.

    The award offers 12 months of free cloud-computing resources to 20 awardees selected from proposals submitted by September 15, 2014.The award offers 12 months of free cloud-computing resources to 20 awardees selected from proposals submitted by September 15, 2014.

    To qualify for the awards program, you must be affiliated with an academic institution or non-profit research laboratory. In addition to individual investigator projects, we are interested in projects that will support access to services and data of value to a collaboration or community.

    Your proposal should not exceed three pages in length. It should include resource requirement estimates (number of core, storage requirements, and so forth) for your project. Apply and learn more about the RFP at Food Resilience Climate Data Initiative.

    We encourage all investigators to join with the USDA and us in an effort to understand the impact of climate change on our food supply.

    Dan Fay, Director for Earth, Energy, and Environment, Microsoft Research

    Learn more

  • Microsoft Research Connections Blog

    From quantum computing to show dogs, Faculty Summit informed and inspired


    Watch Faculty Summit 2014 now on demand

    The fifteenth annual Microsoft Faculty Summit is over, but you can still experience much of it on demand. I was really inspired and energized by the keynotes, session topics, and discussions—especially meeting, talking to, and hearing from researchers in other areas who I don’t normally see at conferences. The Faculty Summit brought together an amazing mix of 500 academics and Microsoft researchers from a wide range of scientific disciplines—from quantum computing to social science.

    Harry Shum, executive vice president of Microsoft’s Technology and Research group, gave the opening keynote speech. Harry's enthusiasm and excitement was palpable as he discussed the future of Microsoft Research. He was inspired by the number of Microsoft and academia collaborations, reminding us how ideas and technologies flow from our industry-academic collaborations.

    There were many great moments, but I was especially delighted by Project Adam, which did a great job of classifying dogs! The goal of Project Adam is a software system that recognizes any object from an image. Researchers have been working on this really hard problem in deep learning for a long time. Johnson Apacible, principal member of tech staff, Microsoft Research, announced Adam has achieved some pretty astonishing results. It is twice as accurate as and 50 times faster than the prior state of the art in literature. The Microsoft team has trained Adam by using DNA models with more than 40 billion connections—and it's still scaling linearly.

    The team took a whimsical approach to the demonstration, asking Adam to "look" at a series of dogs and identify each individual canine's breed. Adam pulled an almost perfect score, but was stumped by the final dog, an Australian Cobberdog, which turned out to be a mix of five breeds. Adam guessed terrier, one of the dog's lineages. Regardless, Adam was more knowledgeable about the dog breeds than many of us in the audience.

    After the keynote, we got into Hot Topics, starting with Doug Burger, who talked about the second age of computing. Doug defined this period as the ending of Moore’s law and explained how this age offers great opportunities for specialization through more efficient and faster computation. Next, Krysta Svore provided a glimpse into quantum simulation and some exciting results that demonstrate the potential for enormous gains in efficiency. Desney Tan then shared his thoughts about individualized healthcare and monitoring and how we (researchers and computer scientists) need to build systems that facilitate real discovery and new approaches by learning from the massive amounts of data that we can capture with billions of body sensors.

    Mary Gray wrapped up Hot Topics by sharing that we are in a revolutionary moment in social science—and in her field of anthropology in particular. "I think we're at the moment where we can think about crowds and the specificity of individual lives in them, and at the same time, use the computational power of computer science to explore patterns that we've never thought about before," Mary said.

    The Keynote and Hot Topics sessions set a high bar for the rest of the summit. Throughout the two days of the summit, a wide variety of topics engaged and delighted us, with more than 20 sessions that examined devices, cloud computing, crowd sourcing, software engineering, quantum computing, how we feel while we program, and much more. I was intrigued by a study shared by Prof. Keith N. Hampton: it indicates that computing is extending our social relationships. Instead of dropping your high school friends when you move away, they remain in your life forever. Is this a good thing or a bad thing? I can’t decide.

    While I watched in person, I was delighted that we were able to share the day one keynotes, Hot Topics session, and engaging interview sessions with the online audience, via a live stream.

    You can watch even more of the Faculty Summit now. We have posted more of the keynotes, sessions, interviews, and demos. Visit the Online Event page to view what you missed, or re-watch your favorite sessions.

    Kathryn S. McKinley, Faculty Summit Co-Chair, Microsoft Research Faculty Summit 2014

    Learn more

  • Microsoft Research Connections Blog

    New hope for people living with paralysis


    When we released the Lab of Things a year ago, we knew that it would benefit researchers experimenting with connected devices in various domains. It has been very gratifying to see how the Lab of Things has helped to accelerate research on helping people with disabilities to live more independent lives.

    Essentially, the Lab of Things is a research platform that enables the deployment of connected devices and sensors at scale. By providing a client-side set of components called HomeOS, the Lab of Things frees researchers from having to develop the complete software stack for deploying their experiments. The HomeOS enables a simple yet powerful connectivity and experiment execution environment. The Lab of Things also comes with a set of cloud services for updating, monitoring, and storage, allowing researchers to scale up deployments and deploy in geographically diverse locations. These features lower the barriers for testing new devices and understanding their behavior in a quick, stable, and repeatable fashion.

    Researchers have been using the Lab of Things to develop new technologies. Professor Nilanjan Banerjee of the University of Maryland at Baltimore County and I recently had the opportunity to describe some of these exciting new technologies during an online webcast of the 2014 Microsoft Research Faculty Summit.

    Professor Banerjee was an early Lab of Things adopter. He approached us shortly after we released it, speaking passionately about developing sensors that could help people with limited mobility lead more independent lives by enabling them to control the environment in their home and workplaces. The Lab of Things and its underlying HomeOS seemed the perfect platform for his project. It would allow him to test his ideas quickly and adapt his design as necessary.

    He started working with Buz Chmielewski, who became a quadriplegic after a surfing accident 25 years earlier. Buz helped Professor Banerjee test his design and arrive at a more usable sensor. After nearly a year, Professor Banerjee and his colleagues had developed a sensor that detects gestures and uses them to activate lights and other appliances in the home. The sensor can be sewn almost anything in the environment—for example, clothing or bedding.

    The sensor design for this project was developed by Professor Ryan Robucci and his team. With the Lab of Things, Professor Robucci was able to develop and test the sensor components quickly without having to develop the accompanying software. Also part of this project was Dr. Sandy McCombe-Waller from the University of Maryland, School of Medicine, who specializes in rehabilitation of people with various forms of injuries and disabilities. She helped with understanding the various types of mobility issues involved, and with the Lab of Things was able to test various designs of the sensor quickly.

    Over the past year, the Lab of Things has also grown in what it offers. Recently we added support for the Arduino hardware prototyping board, opening up the Lab of Things to a whole new world of experimentation with new sensors and devices. The Lab of Things also supports web calls to services such as Weather Underground. All of the apps and drivers are available as sample code for users to adapt.

    Arjmand Samuel, Senior Research Program Manager, Microsoft Research

    Learn more

Page 2 of 111 (332 items) 12345»