This week’s list of data news highlights covers January 23-29, 2016 and includes articles about how activists in Buenos Aires are digitally mapping the city’s slums and a new smartphone-powered device that can identify counterfeit drugs.
AlphaGo, an artificial intelligence (AI) program developed by Google’s AI research division DeepMind to play the complex strategy board game Go, has beaten Fan Hui, Europe’s Go champion. Unlike with chess and other games, many AI researchers have struggled to develop software capable of playing Go better than human experts due to the incredibly complicated strategies involved in the game. DeepMind developed AlphaGo with deep learning algorithms trained on data about how expert players play Go, and then had the system play against itself to further the potential patterns and strategies AlpaGo could recognize and play. AlphaGo’s victory over a human Go expert comes nearly a decade before some AI researchers expected it would be possible for a computer program to do so.
Uber is piloting a program in Texas to use data from Uber drivers’ smartphones to detect signs of dangerous driving so it can crack down on bad drivers. If an Uber user submits a complaint that an Uber driver drove dangerously, Uber will attempt to verify these claims by analyzing data from the driver’s smartphone’s accelerometer and location tracking services, which can indicate dangerous behaviors such as rapid acceleration, hard braking, and frequent lane changes, and discipline the driver accordingly.
Microsoft has made its Computational Network Toolkit (CNTK), its artificial intelligence platform, freely available for the public to use and modify. Microsoft uses CNTK in applications such as Cortana digital assistant for speech recognition and Skype Translate to enable near real-time translation of spoken languages in video calls. Microsoft had initially made CTNK freely available for non-commercial users in April 2015, but by making it fully open source, Microsoft joins the likes of Google and Baidu, which have also made their AI platforms freely available in recent months to encourage its use in academic and commercial applications.
Researchers at the University of California, Berkeley have developed a wrist-worn device that can analyze the chemical makeup of a user’s sweat, which could provide new health and fitness insights. As an individual exercises, sensors in the device track the sodium, potassium, glucose, and lactate concentrations in the wearer’s sweat, and transmit the data, which could indicate factors such as hydration or muscle exhaustion, to a smartphone for analysis. Eventually, the researchers hope to integrate their technology into wearable fitness trackers or other smart devices to give users a more comprehensive view of their health.
An initiative in Buenos Aires called Caminos de la Villa is helping activists improve quality of life for residents of slums, by digitally mapping these neighborhoods for the first time. Residents of Buenos Aires’ slums make up 10 percent of the city’s population, but because most of these neighborhoods do not have official street names or address systems, many cannot reliably receive mail, call an ambulance, or utilize other important municipal services. Caminos de la Villa equipped activists with handheld GPS devices to walk through every street and alleyway in various neighborhoods to digitally record its layout. In the 17 slums Caminos de la Villa has mapped so far, residents can now submit geotagged complaints about the need for municipal services, such as a road in need of repair, to city officials.
India has selected the first 20 cities that will receive funding for its Smart Cities Mission, which aims to develop 100 smart cities in the country by 2022 that utilize information technology to improve infrastructure and enhance municipal services. These 20 cities will receive a total of $7.5 billion to begin developing necessary components of a smart city, including improved public utilities, efficient public transportation systems, and digital infrastructure.
Bay Area Rapid Transit, San Francisco’s public transportation authority, has partnered with transportation planning startup Urban Engines to implement the company’s Space/Time Engine, mapping technology that can help city officials better manage traffic with comprehensive models of how commuters move through the city. The Space/Time Engine allows city officials to monitor traffic patterns using de-identified sensor data about commuters’ movements. In addition, it creates simulations so officials can route public transportation more efficiently during disruptive events such as weather or concerts.
The U.S. Department of Commerce has launched the Commerce Data Usability Project, an initiative to make agency data sets with high economic value more usable by the public. The project, run in partnership with companies including Microsoft, Zillow, and MapBox, provides tutorials on how to analyze and apply particularly valuable data sets. For example, the project launched with tutorials on how analyze the risk of hail damage using severe weather data sets from the National Oceanic and Atmospheric Administration and how to identify communities in need of particular types aid or services using data from the U.S. Census Bureau.
San Jose-based startup Stratio has developed a smartphone-powered device capable of analyzing a pill and comparing this data with a pharmaceutical database to verify whether or not the pill is genuine. The device, called Linksquare, is a miniature spectrometer—a sensor that measures light wavelengths—that analyzes how a pill reflects light, as different chemicals and materials produce different but distinct reflective signatures. If this data differs from the correct reflective signature stored in Stratio’s database, the pill is likely counterfeit. Stratio has demonstrated that LinkSquare is capable of correctly identifying counterfeit Viagra pills, and it is working on expanding its database to analyze other types of drugs.
The U.S. Department of Agriculture (USDA) has selected the winners of the USDA-Microsoft Innovation Challenge, which began in July 2015 as part of the White House’s Climate Data Initiative to improve the resiliency of the U.S. food system with open data. USDA published over 100 years of agriculture and climate data for the challenge so participants could develop new tools and applications that could boost food production. The winning application, called Farm Plenty, allows farmers to analyze local crop trends, understand what crops are increasing in popularity, and monitor crop price fluctuations to help them make more informed decisions about what to plant.