10 Bits: the Data News Hotlist
This week’s list of data news highlights covers March 17 – 23, 2018, and includes articles about how an AI system discovered 6,000 new viruses and a helmet that can monitor brain activity while the wearer moves freely.
Health technology startup Pear Therapeutics has developed a smartphone app called Reset that doctors can prescribe to patients to help manage addiction. Reset allows users to log data about their addiction status, including the strength of cravings and information about factors triggering cravings, such as loneliness or pain, and prompts users to complete 61 included therapy units designed to serve as behavior modification therapy to help them cope with their addiction. Doctors can monitor their patients’ progress through a dashboard. Reset is the first app to be approved for treatment of substance use disorders by the U.S. Food and Drug Administration.
The U.S. Defense Advanced Research Projects Agency (DARPA) has launched a research project with researchers from the University of Wisconsin-Madison and PARC, a research and development company, called D-FOCUS to use AI to generate design concepts for machines such as vehicles and aircraft. D-FOCUS generates alternative versions of existing machine designs that are capable of solving engineering challenges, such as transporting water uphill, to help automate early stages of the design process so that human designers can more easily identify optimal designs.
Researchers at the Department of Energy have used machine learning to identify over 6,000 new species of viruses in the Invoridiae family, of which scientists had previously only identified 100 species. The researchers trained their system on a dataset of 805 genomic sequences of known Inoviridae species and a dataset of 2,000 genomic sequences of a wide variety of bacteria and other viruses to teach it how to distinguish between Inoviridae genetics and those of other organisms. Then, by having this system analyze large genomic datasets, the researchers were able to identify large amounts of previously unknown Invoridae species.
IBM has launched a new service called Watson Assistant that allows companies to tailor IBM’s Watson cognitive computing platform to serve as a voice assistant for any kind of product or service. Companies can train Watson Assistant on their own data to develop relevant voice commands and actions, such as controlling a car’s dashboard or interacting with smart home technologies.
Chinese hospitals and technology companies are increasingly turning to AI-powered health care tools to help alleviate the demand for more doctors, as China only has 1.5 doctors per every 1,000 people, whereas the United States has 2.5 doctors per 1,000 people. 131 companies are developing AI tools for health care and hospitals are deploying systems that can help increase efficiency. For example, a Beijing hospital that sees 10,000 outpatients per day, will begin using AI in April to analyze lung scans to speed the screening process.
A programmer named Alex Bell has developed a machine learning system that can analyze video footage of street traffic and identify when vehicles obstruct bike lanes, which can be dangerous. Bell had his system analyze 10 days of video footage of a block in Harlem provided by the New York City Department of Transportation’s open data portal and found that bike lanes are obstructed 40 percent of the time, and bus stops are obstructed 57 percent of the time.
San Francisco conservation nonprofit Rainforest Connection has developed an inexpensive acoustic monitoring system that can be deployed in a forest to detect potential signs of illegal logging and poaching. Rainforest Connection is also developing a machine learning system with Google’s open source machine learning library TensorFlow to allow the system to more accurately differentiate between these signs, such as chainsaws and gunshots, and other noises.
Researchers at the University of Nottingham and University College London have developed a portal device worn as a helmet capable of sensing a wearer’s brain activity while allowing them to move around freely. Traditionally, monitoring brain activity requires a person to sit or lie motionless while connected to a large magnetoencephalography (MEG) machine, which makes it difficult to analyze brain function related to navigation as well as infants and people with movement disorders. The helmet uses specialized, miniature magnetometers that affix to a person’s head over the target area of their brain to gather brain activity data without the risk of a person’s movement reducing its accuracy.
The World Anti-Doping Agency (WADA) has announced it will pilot AI systems that can flag suspicious activity that could indicate an athlete is illegally doping and improve how WADA officials target testing. WADA gathers large amounts of biological data and other information about athletes, but its efforts to fight doping are limited by a lack of resources. WADA hopes to use AI to help sift through this data to more quickly identify signs of doping as well as identify subtle anomalies that humans would likely miss.
A company called Cogito has developed an AI system that can analyze voice conversations and detect tone and emotion to help call centers better communicate with callers. The system can flag positive or negative changes in emotional state in a person’s voice, allowing call center agents to detect when they may be expressing an inappropriate tone for a particular conversation, such as speaking cheerily when discussing bereavement benefits. Cogito can also track how customers’ tone change in response to different actions.
Image: biology pop.