Lifeboat Foundation

These are just some examples of how easy it is to break the leading pattern-recognition technology in AI, known as deep neural networks (DNNs). These have proved incredibly successful at correctly classifying all kinds of input, including images, speech and data on consumer preferences. They are part of daily life, running everything from automated telephone systems to user recommendations on the streaming service Netflix. Yet making alterations to inputs — in the form of tiny changes that are typically imperceptible to humans — can flummox the best neural networks around.

Artificial-intelligence researchers are trying to fix the flaws of neural networks.

Smile! You’re on camera — or you were at some point in the past few years — and now your face is public domain.

Facial recognition technology is everywhere, and only becoming more pervasive. It’s marketed as a security feature by companies like Apple and Google to prevent strangers from unlocking your iPhone or front door.

It’s also used by government agencies like police departments. More than half of adult Americans’ faces are logged in police databases, according to a study by Georgetown researchers. Facial recognition technology is used by governments across the globe to identify and track dissidents, and has been deployed by police against Hong Kong protesters.

Engineers at the University of California San Diego have developed a way to build soft robots that are compact, portable and multifunctional. The advance was made possible by creating soft, tubular actuators whose movements are electrically controlled, making them easy to integrate with small electronic components.

As a proof of concept, engineers used these new actuators to build a soft, battery-powered robot that can walk untethered on flat surfaces and move objects. They also built a soft gripper that can grasp and pick up small objects.

The team, led by UC San Diego mechanical and aerospace engineering professor Shengqiang Cai, published the work Oct. 11 in Science Advances.

Lex Fridman, a postdoctoral associate at the MIT AgeLab, had a great conversation with Peter Norvig, an American computer scientist, teacher and author.

Vid Robots normally need to be programmed in order to get them to perform a particular task, but they can be coaxed into writing the instructions themselves with the help of machine learning, according to research published in Science.

Engineers at Vicarious AI, a robotics startup based in California, USA, have built what they call a “visual cognitive computer” (VCC), a software platform connected to a camera system and a robot gripper. Given a set of visual clues, the VCC writes a short program of instructions to be followed by the robot so it knows how to move its gripper to do simple tasks.

“Humans are good at inferring the concepts conveyed in a pair of images and then applying them in a completely different setting,” the paper states.

President Nana Addo Dankwa Akufo-Addo on Thursday opened Ghana’s newest medical drone delivery base at Asante Mampong as part of his duty tour of the Ashanti region this week.

This marks the second of what will be four medical drone delivery bases that have been commissioned to help expand access to critical and life-saving medicines for people across Ghana.

Zipline, a California-based robotics company, operates the medical drone delivery service with support from Gavi, the Vaccine Alliance, the Bill & Melinda Gates Foundation, the UPS Foundation and other partners.

3D Printing was one example of how technological advancement made manufacturing accessible to all. Will robots take all the jobs? I doubt it, but technological advancement will make many things inaccessible, accessible to many more than before, by lowering the cost of production. This is but one example.

Better clear out several shelves of storage space, vinylheads, because your record collection is about to expand into infinity. Soon, you’ll be able to get absolutely anything on vinyl. Even better—you’ll be able to make it.

The Phonocut is an at-home vinyl lathe, allowing anyone with a digital audio file and a dream to make a 10-inch record.

A new paper from researchers at the University of Chicago introduces a technique for compiling highly optimized quantum instructions that can be executed on near-term hardware. This technique is particularly well suited to a new class of variational quantum algorithms, which are promising candidates for demonstrating useful quantum speedups. The new work was enabled by uniting ideas across the stack, spanning quantum algorithms, machine learning, compilers, and device physics. The interdisciplinary research was carried out by members of the EPiQC (Enabling Practical-scale Quantum Computation) collaboration, an NSF Expedition in Computing.

Adapting to a New Paradigm for Quantum Algorithms

The original vision for quantum algorithms dates to the early 1980s, when physicist Richard Feynman proposed performing molecular simulations using just thousands of noise-less qubits (quantum bits), a practically impossible task for traditional computers. Other algorithms developed in the 1990s and 2000s demonstrated that thousands of noise-less qubits would also offer dramatic speedups for problems such as database search, integer factoring, and matrix algebra. However, despite recent advances in quantum hardware, these algorithms are still decades away from scalable realizations, because current hardware features noisy qubits.

From facial recognition security to unmanned vehicles, Alibaba’s corporate campus is the office of the future. CNBC’s Uptin Saiidi gets a rare look inside the company’s headquarters in Hangzhou, China, where more than 20,000 employees are based.

–

Continue reading “We went inside Alibaba’s global headquarters | CNBC Reports” »