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

    Software promises timely diagnosis of pneumonia in ICU patients


    The advent of electronic medical records is transforming the care of patients in hospitals, clinics, and doctors’ offices around the world. Offering complete documentation of a patient’s medical history, digitized records are improving the accuracy of diagnoses and the continuity of patient care, which, in turn, means improved patient outcomes.

    Electronic medical records can be especially useful in the diagnosis of pneumonia, which has a nasty habit of appearing after a patient has been hospitalized in the intensive care unit (ICU). Currently, such diagnoses are made by consensus, after a thorough chart review of the patient’s medical tests and clinical notes.

    This means that a physician must comb through hand-written records and copious test results to reach the correct diagnosis—a time-consuming, resource-intensive process. Now, my colleagues and I at the University of Washington, in collaboration with Microsoft researcher Lucy Vanderwende and using Microsoft Research Splat (Statistical Parsing and Linguistic Analysis Toolkit), have created deCIPHER, a project that demonstrates the potential to use natural language processing (NLP) and machine learning to diagnose such critical illnesses automatically from the patient’s electronic medical records.

    We studied the diagnosis of pneumonia in approximately 100 patients being treated in the ICU at Seattle’s Harborview Medical Center. By using the electronic medical records of these patients—whose pneumonia diagnoses had already been established by clinical consensus—we employed state-of-the-art NLP tools from Microsoft to identify the critical clinical information. We then ran that data through a machine-learning framework to see if the software could be trained to correctly diagnose the pneumonia cases based solely on an automated review of the digital medical records. The results were so promising—the software achieved a correct diagnosis with correct time-of-onset for positive cases in 84 percent of the patients—that our clinical collaborators are considering the addition of our pneumonia-detection models to the dashboard they use to monitor ICU patients.

    Meliha Yetisgen, Assistant Professor of Biomedical Informatics, University of Washington

    Related links

  • Microsoft Research Connections Blog

    Microsoft Research – Inria Joint Centre: inventing today, tomorrow’s world


    When world-class research organizations work together on a long-term basis, the whole is greater than the sum of the parts. That premise underlies Microsoft Research’s collaborative projects and joint ventures around the globe, including our recently renewed joint research center with Inria (the French Institute for Research in Computer Science and Automation).

    Since its founding in 2006, the Microsoft Research – Inria Joint Centre has innovatively applied computer science and mathematics to a host of scientific challenges, from formal methods for mathematics to distributed systems and security, computer vision and medical imaging, machine learning and big data, and social networks and privacy.

    From 4D cardiac MR images to mathematical components, researchers gather at the Inria Joint Research Centre

    Microsoft Research – Inria includes 100 researchers overall: 40 permanent researchers from Inria, 30 permanent researchers from Microsoft Research, and 30 non-permanent researchers (interns and postdoctoral and PhD students, representing some 23 nationalities). Today, May 19, the Joint Centre continued its quest to use computing to help solve big problems, hosting an event that reported on the ambitious projects currently underway (see the list later in this blog). The event also featured the following keynotes from some of the world’s foremost computing experts, including Jeanette Wing, corporate vice president of Microsoft Research, who gave an inspiring presentation on how the joint research center is important to science, technology and society.

    Jeanette Wing, corporate vice president Microsoft Research
    Jeanette Wing, corporate vice president of Microsoft Research

    • Thinking For Programmers: Rising Above the Code”: Leslie Lamport, this year’s Turing Award winner and a principal researcher at Microsoft Research Silicon Valley, discussed the need for programmers to create extremely rigorous specifications before coding complex systems, particularly concurrent and distributed systems.
    • “Machine learning for Brain Imaging: from pattern analysis to brain atlases”: Bertrand Thirion, director of research at the Joint Centre, spoke about using machine learning to extract patterns of neurological activity that can lead to a functional atlas of the brain.
    • Formal components for the odd order theorem”: Georges Gonthier, principal researcher and team leader at Inria, focused on how to combine software engineering, programming language, and formal logic techniques to package formal mathematical theories into components that lend themselves to computer-checked formalization of results.
    • Big Learning: New Challenges and Opportunities”: Francis Bach, a senior researcher and team leader at Inria, reviewed recent developments in machine learning—such as improvements in algorithm speed and the use of generalized learning representations—that are tailored to solving modern large-scale problems.

    Georges Gonthier, principal researcher and team leader at InriaGeorges Gonthier, principal researcher and team leader at Inria 

    Bertrand Thirion, director of research at the Joint Centre Bertrand Thirion, director of research at the Joint Centre

    The Joint Centre is currently focusing on the following projects:

    Projects on formal methods and their applications

    • Mathematical Components aims to develop the ability of existing proof assistants, such as Coq, to automatically check difficult proofs in mathematics.
    • Temporal Logic of Actions for Proof System addresses challenges in certifying correct behavior of distributed and concurrent systems, in which there is no certainty as to when distinct components will interact.
    • Secure Computing develops new languages and associated certification tools to prove that implementations of cryptographic protocols are sound, thereby improving the security of Internet transactions.

    Projects on machine learning and big data

    • Large-scale Structured Machine Learning develops new methods for achieving efficient trade-offs between statistical accuracy and computational cost. It also develops algorithms that efficiently trade off exploration with exploitation in active learning scenarios.
    • Z-Cloud Workflows develops solutions for efficiently instantiating workflows in a cloud-computing environment by mapping tasks of the workflow to specific machines. It conjointly optimizes the replication of data within the cloud computing nodes.
    • Interactive Network Visualization develops tools for interacting with and visualizing data that arises from both online social networks and brain imagery, with a particular emphasis on time series.
    • White Box Search-Based Software Engineering uses machine learning to improve software engineering by automatically determining software parameters and assisting developers through the recommendation of code snippets.

    Projects on computer vision and medical imaging

    • Video Understanding aims to extract rich features automatically from large video catalogues, in order to support semantically rich queries when searching such catalogues.
    • Medilearn develops personalized models that assist in the diagnosis and treatment of heart conditions. It also focuses on identification of human brain activation patterns induced by conducting specific cognitive tasks.

    Projects on social networks and privacy

    • Social Information Networks develops efficient recommendation of contacts and contents to users of online social networks. It also addresses the design of reward schemes for incentivizing efficient filtering of information by users.
    • Privacy-Friendly Services and Apps develops means for users to protect their private information, such as geo-localization traces, while preserving the ability of applications to provide value-added services.

    All told, this one-day event captured the essence of the valuable research taking place at the Microsoft Research – Inria Joint Research Centre, and it points out the value of our long-term investments in collaborative ventures.

    Scarlet Schwiderski-Grosche, Senior Research Program Manager, Microsoft Research Connections EMEA

    —Pierre-Louis Xech, Microsoft Research-Inria Joint Centre Deputy Director, Microsoft France


    Learn more

  • Microsoft Research Connections Blog

    The warm heart of Chile embraces TouchDevelop and the Lab of Things


    Chile is a long way from Microsoft Research Redmond, but its bright, inquisitive students and talented, motivated professors share our fascination in the promise of innovative software technologies. Our shared values were on clear display when representatives from Microsoft Research visited the University of Chile and the Pontifical Catholic University of Chile on May 6, 2014. Such visits enable us to check our technologies in new environments, and they always raise interesting new avenues to pursue.

    Professor Sergio Ochoa was our host at the University of Chile, the country’s oldest and largest institution of higher learning, with approximately 38,000 students spread across a full range of academic divisions.  The University of Chile was founded in 1842, the first in the country, and now has students in a full range of faculties and schools. We met faculty who had come from all over the world, enriching the domains and standards of the opportunities for the students.

    At the University of Chile: Prof. Sergio Achoa, Dr. Michael Moskal, Dr. Judith Bishop, Prof. Maria Cecilia Bastarrica, Diego Muñoz, Prof. Alexandre Begel, Prof. Jeremy Barbay, and Dr. Mircea LunguAt the University of Chile, standing, from left: Prof. Sergio Ochoa, Dr. Michal Moskal, Dr. Judith Bishop, Prof. Maria Cecilia Bastarrica, Diego Muñoz, and Prof. Alexandre Begel. Kneeling from left: Prof. Jeremy Barbay and Dr. Mircea Lungu

    While there, we presented a workshop entitled, “TouchDevelop: Create Rich Mobile Cloud Apps on Your Device” to students from the computer sciences department. During the workshop, students peppered us with perceptive questions, particularly about the cloud experience that TouchDevelop offers. In explanation, Michal Moskal, a researcher at Microsoft Research, developed a chat program in TouchDevelop, showing how data can be given a “cloud” tag that makes it updateable by many users simultaneously.

    The obvious follow-up question was whether TouchDevelop could also enable several programmers to work on the same code simultaneously. We believe that teamwork is very important, and we were glad to be able to announce that this capability is being built into TouchDevelop and will be released soon.

    Students work on multiple platforms with TouchDevelop at the University of Chile.
    Students work on multiple platforms with TouchDevelop at the University of Chile.

    At the Pontifical Catholic University of Chile, Professor Andrés Neyem organized a workshop titled, “Lab of Things – Deploying Connected Devices for Research.” At the workshop, Arjmand Samuel and Ratul Mahajan, both from Microsoft Research, talked about the Lab of Things and how this platform could be used to scale up research that relies upon connected devices and sensors. Faculty and students who were present at the workshop raised interesting questions regarding the infrastructure, network protocols, and the security and privacy of data collected as part of such research. Professor Neyem also talked about his research interests in connecting healthcare devices in homes and beyond. Specifically, he showed a demonstration of a pulse-rate monitor that is being developed in collaboration with the university’s School of Nursing, which could be deployed to about 25 homes by using the Lab of Things.

    As eager as the students and faculty were at both universities, we came away just as enthused about possible links with these outstanding institutions. We look forward to working with the University of Chile to improve TouchDevelop and expand its reach, and to collaborating with the Pontifical Catholic University of Chile in deploying the Internet of Things.

    Judith Bishop, Director of Computer Science, Microsoft Research

    Arjmand Samuel, Senior Research Program Manager, Microsoft Research

    Learn more

Page 5 of 109 (327 items) «34567»