This week’s list of data news highlights covers March 14-20, 2020, and includes articles about combating the coronavirus and using AI to create a smart shoe.
Researchers from Intel have designed a new computer chip that helps an algorithm distinguish between different smells using only a single training example for each scent it learns. The chip achieves its efficient structure by mimicking the design of the olfactory bulb in the brain, which plays an essential role in how mammals smell. The chip could be useful for activities such as detecting bombs or dangerous fumes.
Researchers from the University of Chicago have developed DeepScribe, an AI system that can translate tablets from ancient civilizations. The researchers trained the system on 6,000 annotated images of tablets, teaching it to identify cuneiform signs with 80 percent accuracy. The system can automate the translation of highly repetitive parts of the tablets, freeing researchers to analyze more challenging aspects.
Hospitals in China are using wearables and autonomous robots to combat the coronavirus. For example, a field hospital has been using smart bracelets to track the temperature, heart rate, and blood oxygen levels of patients and medical staff to catch early signs of infection. In addition, hospitals in Wuhan are using disinfecting robots, which move around using LiDAR and shine ultraviolet light to decontaminate surfaces, from UVD Robots, a company based in Denmark.
Collective Liberty, a Washington, D.C.-based nonprofit, is using AI to identify criminal networks involved in human trafficking. The organization applies AI to analyze data such as court documents, news reports, and Freedom of Information Act requests to find connections between bank accounts and past prosecutions. The firm’s technology is helping fight trafficking in 204 jurisdictions across 37 states.
Researchers from Nvidia have developed an AI system that helps robotic hands better grasp objects humans hand to them. The researchers trained the system on over 150,000 images of humans handing objects in different ways, teaching the system the best ways to grasp an object given how a human was holding it. The system grasped objects with a 100 percent success rate.
Persivia, a software company based in Massachusetts, has developed an AI-enabled tool that can predict which patients likely have COVID-19. The tool uses machine learning to analyze data such as vitals signs and social and travel histories in patient notes to recommend a course of action, such as to test an individual immediately. Prima Health Care Service, an organization that owns dozens of hospitals in the United States, is using the tool to ensure it does not overlook any potential coronavirus cases.
Researchers from MIT have developed a sensor that could prevent food waste by monitoring fruits’ and vegetables’ levels of ethylene, an odorless gas that plants emit when they ripen. The sensors work by using palladium, a metal, to add oxygen to ethylene, which causes the metal to pass electrons to the sensor, making it more conductive. The sensors measure this change in current flow to detect the presence of ethylene.
Nokia has developed an IoT system that can detect and provide early warnings for landslides. The system can give managers on highways real-time data on changes in ground conditions and stability for highway slopes and alert the managers to potential landslides. Nokia tested the system in the Guangxi province in China, where it helped prevent harm to drivers following a landslide in March 2019.
Microsoft has developed an AI-enabled system that helps New York fashion company Bode, which makes new garments out of aged quilts, to automatically describe the quilts to customers. Microsoft trains the system on hundreds of images of roughly 30 different examples of a quilt to allow it to identify and describe key design details. For example, the system could identify that the quilt features a 1950s “Log Cabin” motif.
Researchers from the Stevens Institute of Technology in New Jersey have developed an AI-enabled smart insole for shoes that allows for portable gait-analysis. The sole uses accelerometers and gyroscopes to monitor its movement and orientation and force sensors to detect plantar pressure. The sole can help researchers measure walking function in patients with movement disorders or injuries without requiring the patients to come to or stay in a lab or medical environment.
Image: Sheri Terris