Weekly News Boston

Published on March 11th, 2016 | by Joshua New

0

10 Bits: the Data News Hotlist

This week’s list of data news highlights covers March 5-11, 2016 and includes articles about an artificial intelligence program that is beating Go grand champions and a new browser extension that encourages scientists to share their data.

1. Building Economic Opportunity with Open Data

The Obama Administration has announced “The Opportunity Project,” an initiative to provide civic technologists, businesses, and local and state policymakers with open data and tools to increase economic mobility. The Census Bureau has launched the website opportunity.census.gov to house useful case studies, tools, and other resources, and it has also released a new module for its City Software Development Kit (CitySDK) to make it easier to build tools with open data. Federal agencies will supply data that can increase economic opportunity for populations in need, such as unemployed or underemployed workers, veterans, and citizens returning to the country that face difficulty reentering the workforce.  

2. Identifying Terrorists by Analyzing Their Hand Signs

Researchers at Mu’tah University in Jordan have developed a machine learning algorithm that can detect the subtle differences in hand geometries in images of people making “V for victory” hand sign commonly used by terrorists, which could help identify terrorists wearing scarves or hoods to hide their identity. The researchers trained their algorithm on 500 images of 50 different people’s hand signs, teaching it to analyze differences in finger size, the angle between the fingers, and other traits. In testing, the algorithm can correctly guess the identity of someone making the hand sign with over 90 percent accuracy.  

3. Artificial Intelligence Masters Go

AlphaGo, Google’s artificial intelligence Go program, beat Go grandmaster Lee Sedol three times in a best-of-five series of matches, before losing the fourth match. Go has more potential moves than there are atoms in the universe, so human players often rely heavily on intuition, rather than calculation, to choose their moves, making the game particularly challenging for computers to play at high skill levels. The researchers that developed AlphaGo trained its neural networks with thousands of human moves and then had the program play against itself—a technique called reinforcement learning, to expose it to new move combinations and increase its skill level.

4. Mapping Air Quality with the Internet of Things

Researchers at the University of Texas, Dallas are developing the Geolocated Allergen Sensing Platform (GASP), which will analyze data from air-quality sensors deployed throughout a city to map the concentration of allergens and pollutants in real time. The sensors will also collect other relevant data, such as temperature, air pressure, and humidity, with the goal of helping people sensitive to air quality, such as asthma and allergy sufferers, better manage their health. Over the next year, the GASP project will deploy sensors in lampposts throughout the university’s campus as well as Chattanooga, Tennessee.

5. Making Genetic Data Publicly Available

Genetic testing company Ambry Genetics has announced it will publish genetic data of 10,000 of its customers who have or have had breast or ovarian cancer to support precision medicine research. The data will only contain aggregate sequencing information about specific genetic variants related to these cancers, so it is not personally identifiable. Other genetic testing companies such as 23andMe offer similar data to pharmaceutical companies working on precision treatments, but Ambry is the first to make this data freely available.

6. Building Glasses that Can Focus Themselves

Israeli startup Deep Optics is developing eyeglasses that can automatically adjust their strength and focus themselves by analyzing data from sensors monitoring where a wearer’s eyes are looking. A small processor in the glasses analyzes data from these sensors to estimate what the wearer is trying to focus on and generates an electrical current to change the refractive index of a liquid-crystal layer on the lenses. Deep Optics is developing this technology to help people with vision disorders such as presbyopia, a common condition in people over 40, which makes it difficult to focus on close-up objects.

7. Mapping Boston’s Energy Demand

Researchers at the Massachusetts Institute of Technology (MIT) have developed a model to map and predict the electricity demand of the city of Boston—information that could help city officials save between $600 million and $1.7 billion over the next 25 years as the city works to improve the efficiency of its power grid. Rather than searching for and analyzing data about the 100,000 buildings in Boston to estimate their energy consumption, the researchers instead built a model that could predict the energy consumption of each building for any given time of day or date, with 94 percent accuracy. With this model, city officials can make more informed decisions about infrastructure planning and reduce strain on the power grid.

8. Using Algorithms to Predict Police Misconduct

Researchers at the University of Chicago have developed an early warning system for police misconduct using data about their interactions with members of the public. The researchers developed their algorithm by analyzing data on a decade of police interactions provided by police officials in Charlotte, North Carolina, and incorporating insights the team learned from police ride alongs, focus groups, and interviews. The researchers were able to identify several new factors that could increase the risk of police misconduct, such as how stressful incidents early in the day, like responding to a suicide or domestic violence call, increase the risk of adverse interactions later in the day. The researchers tested their algorithm on historical data and found that it was able to flag more officers that went on to commit misconduct but flagged 51 percent fewer officers overall, compared to an older system developed by the Chicago Police Department.

9. Tracking the Spread of Zika with Crowdsourced Data

Flu Near You, a free website that allows users to anonymously report health information so it can monitor the spread of the flu, has expanded its data collection tools to include symptoms of the Zika virus, such as joint pain and yellow skin or eyes. Flu Near You maps reports of these symptoms to augment offical disease surveillance programs, such as those conducted by the U.S. Centers for Disease Control, and help public health officials better detect outbreaks and track the spread of the disease. The symptoms of Zika will also help Flu Near You monitor dengue fever and the chikungunya virus, which have similar symptoms.

10. Convincing Scientists to Share Their Data

A volunteer team supported by the nonprofit Center for Open Science has developed the Open Data Button, a web browser add-on that allows users reading research papers online to send requests for the underlying data to the paper’s author, to encourage scientific data sharing. All data requests are also posted on the Open Data button website so others can comment or add their support for the publication of a particular data set. The Open Data Button is an extension of the Open Access Button project, which developed a similar tool to increase public access to research papers behind pay walls.

Image: skeeze.

 

Tags: , , , , , , ,


About the Author

Joshua New is a policy analyst at the Center for Data Innovation. He has a background in government affairs, policy, and communication. Prior to joining the Center for Data Innovation, Joshua graduated from American University with degrees in C.L.E.G. (Communication, Legal Institutions, Economics, and Government) and Public Communication. His research focuses on methods of promoting innovative and emerging technologies as a means of improving the economy and quality of life. Follow Joshua on Twitter @Josh_A_New.



Back to Top ↑

Show Buttons
Hide Buttons