This week’s list of data news highlights covers August 22-28, 2015 and includes articles about a computer program can diagnose psychosis by listening to someone talk and how teaching robots to make pizza might help researchers develop better robotic problems solving.
Uber has announced a partnership with the University of Arizona’s College of Optical Science to research and develop imaging technology to improve the mapping and safety capabilities of self-driving cars. In conjunction with the partnership, Arizona Governor Doug Ducey issued an executive order to allow for testing of self-driving cars in the state.
The United Nations Food and Agriculture Organization (FAO), the group’s agriculture agency, has announced the development of a new database designed to help water-scarce countries make better use of their limited water for agricultural purposes. The free-to-access database will contain satellite imagery and analysis that should help countries use water and arable land more productively. The database will be available to any country that wishes to access it, but it will focus primarily on historically water-scarce countries in West Asia and North Africa.
Researchers at Columbia University, the New York State Psychiatric Institute, and the IBM T.J. Watson Research Center, have developed software that can correctly differentiate between patients at risk of developing psychosis and those not at risk, with 100 percent accuracy in a recent study, just by analyzing their speech. The speech analysis, which proved more accurate than other screening methods such as neuroimaging and brain activity analysis, focused on the complexity and coherence of patient speech, which can be a strong indicator of schizophrenia, a symptom of psychosis. Researchers hope their method could eventually be used to signal the need for early intervention and treatment for patients at risk of developing schizophrenia.
The National Oceanic and Atmospheric Administration (NOAA) and the National Aeronautics and Space Administration (NASA) have partnered to equip a drone with weather data sensors to improve storm forecasting, as drones can collect real-time data in weather conditions that manned aircraft cannot. The project will begin with Tropical Storm Erika, which is currently headed towards the southeast United States. The drone will continuously collect data on moisture, wind speed, and wind direction from above the storm, too high an altitude for manned aircraft, or in front of the storm. NASA and NOAA hope this data could drastically improve forecast models.
A company called Now-Cast Data Corporation has created software capable of constantly generating forecasts for the economy by continuously analyzing 3,600 indicators that include social media data and the prevalence of certain Internet searches. The program, called Xcalibur, calibrates its predictions based on this data to give users a real-time view of the economy, which Now-Cast expects could help investors and policymakers make better decisions, as they traditionally rely on estimates that take weeks or months to produce.
Google’s Government Innovation Lab, launched in April 2015 to help local governments make better use of technology, has helped California’s Kern County produce its first two prototypes to help solve problems facing the county government. The first, called the Virtual Resource Library, is an online hub designed to connect citizens with government services, such as education, training, and even job hunting services. The second prototype is an app designed to pull county government data into a centralized database for analysis purposes. Kern County’s first project with this app will be analyzing community health data to facilitate preventative care.
Facebook has begun testing an artificial intelligence service called M to serve as a digital assistant for users of its Messenger app. Facebook has rolled out M to small test groups of people to help fine tune its algorithms. M will attempt to automatically complete tasks users ask of it, such as providing weather updates or setting reminders. In addition, Facebook staff will help train the software to deliver more useful results by providing it answers.
Researchers partaking in a European project called RoboHow are attempting to teach robot-controlling software how to solve unfamiliar tasks by providing it instructions on how to prepare food from WikiHow, the open source “how-to” website. The goal of the project is to develop software that can figure out how to solve a task without having to program a robot to perform specific actions. Once a robot successfully identifies how a set of instructions relates to a task, the researchers share this information to a database called Open Ease, an open source database for artificial intelligence knowledge.
U.S. Chief Data Scientist DJ Patil and Stephanie Devaney, project manager for the Precision Medicine Initiative, have issued a call for feedback on how to improve personalized medicine, such as ideas about how to improve data-sharing between healthcare organizations and how to include underserved populations in research. The Precision Medicine Initiative will collect ideas on this topic until September 21, 2015.
An initiative called Digital Matatus, made up of researchers from Columbia University, the Massachusetts Institute of Technology, the University of Nairobi, and a design firm called Groupshot, have successfully digitized Nairobi’s bus system. Much of Nairobi’s population relies on privately-owned miniature buses with no formal coordinated system of bus routes or schedules. University students working with Digital Matatus plotted the piecemeal system of buses over four months by logging their routes with location tracking data from their smartphones. As a result, Nairobi residents can now view the bus system using Google Maps.
Image: Hansueli Krapf.