Researchers show how they can reverse engineer and reconstruct someone else’s machine learning engine—using machine learning.
How to Steal an AI
Posted in robotics/AI
Posted in robotics/AI
IBM today announced today Watson -based “Project DataWorks,” the first cloud-based data and analytics platform to integrate all types of data and enable AI-powered decision-making.
Project DataWorks is designed to make it simple for business leaders and data professionals to collect, organize, govern, and secure data, and become a “cognitive business.”
As the saying goes, “If you want something done right, you gotta do it yourself,” and it seems that you’ll soon be able to get a lot more done using artificially intelligent, high-tech exoskeleton Kindred. It’s the product of a startup created by quantum computing company D-Wave’s founder Geordie Rose, and according to the venture capital firm funding Kindred, the device “uses AI-driven robotics so that one human worker can do the work of four.”
Based on a patent application, the wearable system is envisioned as a 1.2-meter tall humanoid that may be covered with synthetic skin. It will include a head-mounted display and an exo-suit of sensors and actuators that carries out everyday tasks.
Essentially, it looks something like Spider-Man’s Doctor Octopus on the outside, but on the inside, Kindred utilizes quantum computation, a way of information processing and storage that is much faster and more powerful than that used by conventional computers. Data “learned” by the suit can be taught to other robots, allowing those robots to then perform the tasks autonomously.
Phototactic behaviour directs some bacteria towards light and others into darkness: This enables them to utilize solar energy as efficiently as possible for their metabolism, or, otherwise, protects them from excessive light intensity. A team of researchers headed by Clemens Bechinger from the Max Planck Institute for Intelligent Systems and the University of Stuttgart, as well as colleagues from the University of Düsseldorf have now found a surprisingly simple way to direct synthetic microswimmers towards light or darkness. Their findings could eventually lead to minuscule robots that seek out and treat lesions in the human body.
Posted in robotics/AI
A new system called HeroSurg, developed by researchers at Deakin and Harvard Universities, is set to increase what surgeons can achieve via robotic surgery, using a haptic feedback system to provide a sense of touch. It also brings other improvements over existing tech, such as collision avoidance, to make robotic surgery safer and more accurate.
Robotic surgery, wherein human-controlled robots perform delicate surgical tasks, has been around for a while. One great example of the tech is the da Vinci robotic surgical system from Intuitive Surgical – a setup made up of numerous robotic arms, a console to operate the instruments, and an imaging system that shows the surgeon what’s happening in real time. In 2008, Professor Suren Krishnan, a member of the team behind HeroSurg, became the first surgeon to perform ear, throat and nose operations using the da Vinci robotic surgical system.
Since then, we’ve seen numerous breakthroughs, including improvements to the original da Vinci system, and other robots emerging capable of achieving impressive tasks, such as performing surgery on a beating heart, or successfully stitching soft tissue.
The exploration of Europa begins under the ice in Antarctica.
That’s where a team of researchers, led by the Georgia Institute of Technology (Georgia Tech), has been testing a variety of robotic subs in recent years to learn about what technologies will work best when NASA eventually launches a mission to Jupiter’s icy moon.
“I really want us to go down through the ice on Europa. I want to explore what’s down there,” says Britney Schmidt, assistant professor at the School of Earth and Atmospheric Sciences at Georgia Tech and principal investigator for the NASA-funded project called SIMPLE, for Sub-ice Investigation of Marine and Planetary-analog Ecosystems.