10 Bits: the Data News Hotlist
This week’s list of data news highlights covers September 17-23, 2016 and includes articles about how ranchers are using the Internet of Things to keep cows healthy and a new open police data tool for California.
Microsoft has announced a series of new initiatives to use artificial intelligence to fight cancer. One of the initiatives, called Project Hanover, will develop machine learning software that can process millions of scientific research papers about cancer to help oncologists develop personalized cancer treatments. Another initiative involves using AI to help radiologists quickly and accurately analyze radiological scans to track disease progression. A third initiative uses a tool called the Bio Model Analyzer to detect cancer earlier and simulate how different types of leukemia will respond to different drugs.
The Amsterdam Institute for Metropolitan Solutions has launched a five-year research initiative to test the use of a specialized autonomous boat called Roboat in Amsterdam’s canals. Researchers will begin testing Roboat, designed like a rectangular floating platform, in 2017, and will experiment with commercial and public applications, such as ferrying goods and people, as well as coordinating with each other to form temporary bridges across canals.
The White House has announced a series of new commitments to support the Open Government Partnership, a coalition of 70 countries committed to increasing accountability, transparency, and citizen participation in government through technology and innovation. The new commitments include improving open contracting data standards, increasing transparency in extractive industries, expanding access to scientific data about climate and arctic issues, and developing a national data sharing platform for projects related to the United Nation’s Sustainable Development Goals. The new commitments expand upon the United State’s third Open Government National Action Plan, which the White House launched in 2015.
Researchers at the Massachusetts Institute of Technology have developed a system called EQ-Radio that uses radar to recognize a person’s emotions by detecting changes in his or her biometric activity. EQ-Radio emits low-power wireless signals that bounce off of people and an algorithm then analyzes subtle changes in this data caused by a person’s changing breathing patterns or heart rate. By associating certain changes in this vital sign data with different emotions, EQ-Radio can determine if a person is joyful, sad, angry, or content with 87 percent accuracy, which researchers expect could be useful information in environments such as hospitals and smart homes.
Cattle ranchers are increasingly using networked sensors and analytics to help manage their herds and reveal information that could help prevent sickness. By tagging cows with location sensors, ranchers can monitor their movement and feeding patterns to detect unusual activity, such as a cow not eating, that could indicate a cow is getting sick and take preventative action. Developers of these monitoring systems estimate that this approach can reduce the amount of antibiotics and medication ranchers use to keep cows healthy, cutting medication costs by 15 percent per cow and reducing the number of cows that die from illness.
Jigsaw, a subsidiary of Google that develops tools to promote freedom of expression and combat extremism online, has developed an artificial intelligence tool called Conversation AI designed to detect and filter out abusive language online to combat online harassment. Jigsaw trained Conversation AI on 17 million comments on New York Times articles and 13 thousand discussions on Wikipedia pages to identify abusive language with 92 percent accuracy. Jigsaw will make Conversation AI available as open source to allow web developers to implement it for their sites as a filter that blocks abusive language in comments.
San Francisco startup Trace Genomics has launched a genomics service that allows farmers to send in soil samples to identify any potential diseases that could harm their crops. Using genetic analysis, Trace Genomics can identify bacteria, viruses, and fungi in soil samples responsible for 30 different diseases that affect strawberries and lettuce, allowing farmers to intervene and treat their crops accordingly. As Trace Genomics accumulates more data, it plans to expand the number of pathogens it can identify and offer more tailored recommendations for farmers.
Analytics startup Bayes Impact, working with the California Department of Justice, has launched a data gathering platform called Ursus to improve how California police departments report instances when a police officer uses force. All police departments in the state are required to use Ursus, which is available for departments to use for free, to log data about every instance of police use of force, and beginning in 2017, Ursus data will be publicly available as open data on California’s Open Justice Portal.
Pennsylvania State University and nonprofit AI research group OpenAI have developed a tool called cleverhans to help AI researchers test how vulnerable their algorithms are to adversarial examples—malicious code specifically designed to trick AI systems. For example, research has shown that AI smartphone assistants such as Apple’s Siri can be tricked by audio recordings, which sound unintelligible to humans, to visit malicious websites. Cleverhans provides researchers with a collection of adversarial examples, defenses against them, and tutorials on how to implement the defenses. Cleverhans is freely available as open source and can be updated to identify new types of adversarial examples.
The U.S. Department of Transportation (DOT) has published proposed guidelines for national oversight of self-driving cars designed to establish safety standards, promote the development of uniform policies, and clarify how regulations would apply to autonomous vehicles without stifling their development. The guidelines include a 15-point safety standard to address issues such as how automakers will protect occupants and communicate with other cars on the road, and a model state policy that fits into a consistent national regulatory framework for autonomous vehicles.
Image: U.S. Department of Agriculture.