This week’s list of data news highlights covers December 7-13 and includes articles on Facebook’s new A.I. lab and a proposed national drug monitoring program to identify patients with high risk of prescription drug abuse.
As part of an initiative to persuade young people to adopt healthy lifestyles, the European Union will fund the development and deployment of sensors that measure how quickly people eat. The sensors, which will be tested on around 200 secondary school students in Sweden and another set in the Netherlands, comprise a scale that connects to a smartphone or laptop; the sensor measures how quickly the weight of the food decreases, indicating how quickly an individual is eating.
The American College of Physicians has released a policy statement with new recommendations for preventing prescription drug abuse in patients, including the formation of a national monitoring program to identify patients at risk for significant drug abuse. This program would allow prescribers and dispensers to check databases in their own and neighboring states before writing prescriptions for substances with high abuse potential.
Facebook has announced the creation of a new artificial intelligence lab, and it has tapped New York University’s Dr. Yann LeCun to direct research. Professor LeCun, one of the world’s top machine learning researchers, is best known for his early work on automated handwriting recognition and reading. He is a pioneer of neural networks, a set of methods that have come to be known as “deep learning” and are increasingly popular in complex machine learning applications.
Popular song-finding app company Shazam is using the data from its users’ searches to predict what musicians will hit the big time next year. The company combines critics’ reviews with the number of people who have used Shazam to find a given song to find artists who are generating the most interest, whether positive or negative. Next year’s predictions include Queens rapper Action Bronson and singer/actress Lucy Hale.
One of the lesser-discussed features of the congressional budget proposal is a provision that would restrict access to a database known as the “Death Master File,” containing names, Social Security numbers and birth and death dates of the deceased. The database was created to help financial institutions scan for fraudulent activity, but its ease of access has actually had the opposite effect, with would-be identity thieves gaining easy access at low cost. Under the budget proposal, information about the deceased would not appear on the database until three years after death, except among a small number of certified organizations.
Organizers of the international ICEGRAV project are using low-flying planes to study the ice sheets of Antarctica in greater detail than any satellite can. A recent data release tracks variations in gravity, magnetism and ice thickness across a region known as the Recovery Catchment, a largely unexplored region that is a crucial object of study for predicting sea level rise. The region is currently stable, but scientists are studying the new data to anticipate any vulnerabilities that could lead to global sea level rise.
Computer scientists at Harvard and the Wyss institute have published research showing that an important class of machine learning algorithms can be implemented using chemical reactions. In the longer term, the researchers hope their work could herald “smart drugs” that can automatically detect, diagnose and treat certain diseases.
As part of the Computer Science Education Week, President Obama delivered a brief address this week encouraging every American to learn to program. The President said that greater computer science literacy was important not only for individuals in the job market, but also for the American economy. Nonprofit computer science education organization Code.org organized the week’s activities, which also include one-hour introductory programming lessons hosted by various groups, including Apple.
HP and nonprofit environmental organization Conservation International have partnered on a system that applies large-scale data analysis to images from the rain forest to predict changes in animal populations. The system showed that several species were at greater risk than scientists realized, including species of anteater, armadillo and wild boar.
International auditors are increasingly aware of the potential of data analytics to aid their work, according to a PricewaterhouseCoopers survey that was delivered by webinar this week. But even though 85 percent of respondents said data analytics is important to strengthening audit coverage, only 31 percent said they were using analytics regularly.
Photo: Flickr user LEOL30