There are many pain points when it comes to the coronavirus, officially known as COVID-19. One of them is how exactly to test people for it when the necessary testing kits are in short supply. One possible solution could be to allow artificial intelligence to scrutinize chest X-rays of patients’ lungs to spot signs of potential coronavirus-caused lung damage.
That’s the basis for several exciting and promising attempts to develop a neural network that could be used to give a strong indication of whether or not a patient likely has COVID-19. Researchers at Chinese medical company Infervision recently teamed up with Wuhan Tongji Hospital in China to develop a COVID-19 diagnostic tool. It is reportedly now being used as a screening tool at the Campus Bio-Medico University Hospital in Rome, Italy.
Meanwhile, other researchers from the University of Waterloo in Ontario, Canada, and Canadian A.I. firm DarwinAI this week announced a new open-access neural net that’s open to the public. The neural net was announced at MIT Technology Review’s EmTech Digital event by DarwinAI CEO Sheldon Fernandez. Called COVID-Net, it’s intended as a tool that could be used for similar screening — and is open for further testing by researchers around the world, who may soon be able to deploy it as a much-needed public health solution.
“We carried [out the A.I.’s] training on a dataset made up of 5,941 posteroanterior chest radiography images, across 2,839 patient cases, from two-open access data repositories,” Alexander Wong, one of the researchers on the project, told Digital Trends. “So far, the sensitivity to COVID-19 cases is quite good. However, the data on COVID-19 cases is still limited and we are continuing to improve the COVID-Net model as more data comes in over time.”
This is the problem that any A.I. researchers are likely to run into. Simply put, there’s still much to learn about COVID-19, which can make developing tools for recognizing it (and, in this case, distinguishing it from other maladies of the lung) difficult. That is why the idea of a publicly available — and publicly scrutable — system is so promising.
“[COVID-Net] is currently not used by patients,” Wong said. “But we are continuing to work hard on improving the results, and invite clinicians and clinical institutes and organizations to use it, give feedback, [and] contribute data so we can accelerate its readiness for clinical deployment. Right now, everything is available to the global community, so hopefully this accelerates progress and advances in this area.”
A.I. researchers are always talking about wanting to solve big problems. Right now, this is one of the biggest that there is.
San Francisco startup Dimensional advertised for people to walk robot dogs through city streets. The New York Times reported that the job could include carrying a roughly 33-pound robot back to the office if it overheated. A lift home is an awkward requirement for a labor-saving machine.
Keeping the hardware running is only one challenge. Anthropic's robotics tests found that language models mostly failed when controlling joints directly, but did better when directing trained controllers. Giving an AI agent a body doesn't teach it to handle ordinary household objects.
Soon, a mattress protector may be more than a bed cover. Australian medtech company Sleeptite and researchers at RMIT University have received a US patent for FlexiSense, a smart bedding technology that tracks body position and posture to help caregivers identify when older residents may be at risk of falling.
The technology is being developed for Sleeptite’s in-bed monitoring system, REMi, and was recently on display at The MedTech Conference in Boston. A peer-reviewed study in Flexible and Printed Electronics says the sensors can also be made at scale without sacrificing performance or durability.
AI companies spend enormous amounts of time convincing us their technology is safe and secure. But what happens when those same companies become targets of increasingly sophisticated AI-powered attacks? Anthropic, the company behind Claude, is now facing that question. According to a Washington Post report, cybersecurity startup Cathedral discovered vulnerabilities that let its researchers access sensitive information in Anthropic's internal systems.
The researchers used AI agents to identify potential entry points in less than 12 hours. Interestingly, they didn't need to crack Claude itself. Instead, they exploited weaknesses in third-party business applications that Anthropic uses for everyday operations. Anthropic's security testing program disclosed the findings, which allows outside researchers to identify and report vulnerabilities.