This week’s list of data news highlights covers June 18-24, 2016 and includes articles about a smartphone app that can warn users about dangers drivers and the winner of the U.S. Department of Transportation’s Smart City Challenge.
The U.S. Defence Advanced Research Projects Agency (DARPA) has launched the Data-Driven Discovery of Models (D3M) program to develop software tools that can make it easier for workers without data science expertise to work with data. D3M will focus on creating algorithms that automate portions of the data modeling process so that subject matter experts can create complex empirical models without additional training. DARPA created D3M to help ensure that employers can take advantage of the benefits of data despite a substantial and growing global shortage of workers with data science skills.
Startup Nexar has released a smartphone app that serves as a “dashcam”—a camera that records through a vehicle’s windshield—that will soon begin using the cameras and other sensors on phones to detect dangerous driving on the road and identify vehicles known to have bad drivers. The app records the license plates of vehicles that exhibit dangerous behavior in front of the user, such as cutting off a user’s car or braking suddenly, to develop driving scores for each vehicle. Later this year, Nexar will update its app to automatically warn drivers when it detects a vehicle with a low score nearby on the road so that they can take extra precautions.
Researchers at the Massachusetts Institute of Technology have developed a deep learning algorithm that can predict what will happen in a scene from a television show based on still images. The researchers trained their algorithm with 600 hours of unlabeled clips from popular shows and had it guess if characters in a scene would hug, kiss, shake hands, or high-five based on the initial frames of the scene. The algorithm was correct 43 percent of the time, and though human subjects were correct 71 percent of the time, the algorithm was substantially more accurate than an untrained control algorithm.
The White House has announced the creation of the Smart Manufacturing Innovation Institute in Los Angeles, which will bring together industry, academic, and non-profit leaders to advance research into connected sensors and digital tools that can substantially improve manufacturing efficiency. The Smart Manufacturing Innovation Institute will receive $70 million from the Department of Energy and an additional $70 million from industry groups and state entities. The institute is the ninth smart manufacturing hub established as part of the Obama Administration’s National Network for Manufacturing Innovation, which aims to boost the competitiveness of the U.S. manufacturing sector with new technologies.
The U.S. State Department has announced it will deploy an analytics platform and install smart electricity meter networks in 22,000 of its buildings around the world to monitor their energy use in real time and improve their efficiency. By the end of 2017, the State Department will install the smart meters in 200 of its 275 posts, each of which which can span hundreds of office buildings, apartments, and large facilities, and it will fully deploy the technology to all of its posts by 2020. By tracking the energy use of all of its buildings in real time, the State Department will be able to identify opportunities to reduce energy consumption and costs.
Researchers at the Austrian Academy of Sciences have successfully simulated the creation of antimatter by using a quantum computer, which can perform the incredibly complex calculations that normal computers typically cannot solve. Regular computers perform calculations with bits representing either “0” or “1,” but a quantum computer uses qubits, which can represent “0,” “1,” or a superposition of both values. Though regular computers can perform this particular simulation, the researchers’ experiment demonstrates that a more powerful quantum computer could open the door for scientists to tackle new and challenging types of physics problems.
The U.S. Department of Transportation (DOT) has selected Columbus, Ohio as the winner of its Smart City Challenge and will award the city with $50 million to develop the United States’ first large-scale smart city pilot project. Columbus will use the funds, as well as an additional $90 million from local industry, to implement a variety of Internet of Things-based projects designed to increase economic opportunity, improve logistics, making transportation more sustainable, and connecting visitors and residents. Projects include deploying connected and autonomous vehicles for public transportation, developing software tools to reduce congestion, and building a “smart corridor” of connected infrastructure and public transportation.
The U.S. General Services Administration (GSA) has implemented the Transactional Data Reporting Rule which requires federal contractors to report detailed data about products and services they provide to the government to improve price transparency and help federal agencies make more informed purchasing decisions. Under the new rule, contractors must electronically report data on all purchases through GSA contracts, including information on price, quantity, and product specifications. GSA expects the rule to save the government $29 million per year by making procurements easier to track.
China’s National Supercomputing Center has unveiled its new Sunway TaihuLight supercomputer, with a computing speed of 93 petaflops per second, meaning it can perform 93 quadrillion calculations per second, which makes it the most powerful supercomputer in the world. With the addition of this supercomputer, China now surpasses the United States as the world leader in supercomputing capacity.
The French government has launched a genomics and personalized medicine program called France Genomic Medicine 2025, focused on improving disease diagnostics and prevention in the country. The French government and industry will invest a total of €670 million ($761 million) to develop genetic sequencing platforms, establish two national centers for data analysis and genomic expertise, and foster public-private partnerships to focus initially on diabetes, cancer, and rare diseases, and expanding to common diseases in 2020. France Genomic Medicine 2025 will establish a government strategic committee to oversee implementation of the program and ensure it meets several objectives, including sequencing 235,000 genomes per year by 2020 and developing the necessary technological infrastructure to support precision medicine.
Image: John Mueller.