Profile cover photo
Profile photo
Wayne Radinsky
Software Design Engineer
Software Design Engineer


Post has attachment
A "robot" made of a single strand of DNA, "can autonomously 'walk' around a surface, pick up certain molecules and drop them off in designated locations." "The researchers constructed three basic building blocks that could be used to assemble a DNA robot: a 'leg' with two 'feet' for walking, an 'arm' and 'hand' for picking up cargo, and a segment that can recognize a specific drop-off point and signal to the hand to release its cargo. Each of these components is made of just a few nucleotides within a single strand of DNA."

"The Lulu Qian group built a robot that could explore a molecular surface, pick up two different molecules -- a fluorescent yellow dye and a fluorescent pink dye -- and then distribute them to two distinct regions on the surface. Using fluorescent molecules enabled the researchers to see if the molecules ended up in their intended locations. The robot successfully sorted six scattered molecules, three pink and three yellow, into their correct places in 24 hours. Adding more robots to the surface shortened the time it took to complete the task."

Post has attachment
Optics rather than electronics switch a nanometer-thick thin film device from completely dark to completely transparent at a speed of trillionths of a second.

"Electrons spinning around inside devices like those used in telecommunications equipment have a speed limit due to a slow charging rate and poor heat dissipation, so if significantly faster operation is the goal, electrons might have to give way to photons."

"The technique uses two laser beams, one carrying the information and the second switching the device on and off. The switching beam uses photons to heat up electrons inside semiconductors to temperatures of a few thousand degrees Fahrenheit, which doesn't cause the sample to get that hot but dramatically changes the material's optical properties. The material also relaxes at terahertz speeds, in a few hundred femtoseconds or in less than one trillionth of a second."

"Sandia researchers turn the optical switch on and off by creating something called a plasmonic cavity, which confines light within a few tens of nanometers, and significantly boosts light-matter interaction. By using a special plasmonic material, doped cadmium oxide from North Carolina State, they built a high-quality plasmonic cavity. Heating up electrons in the doped cadmium oxide drastically modifies the opto-electrical properties of the plasmonics cavity, modulating the intensity of the reflected light."

"Traditional plasmonic materials like gold or silver are barely sensitive to the optical control beam. Shining a beam onto them doesn't change their properties from light to dark or vice versa. The optical control beam, however, alters the doped cadmium oxide cavity very rapidly, controlling its optical properties like an on-off switch."

Post has attachment
"A robot dentist has carried out the first successful autonomous implant surgery by fitting two new teeth into a woman's mouth, mainland media has reported. Although there were human medical staff present during the operation, they did not play an active role while it was being carried out. The one-hour procedure took place in Xian, Shaanxi, on Saturday, according to Science and Technology Daily. The implants were fitted to within a margin of error of 0.2-0.3mm, reaching the required standard for this kind of operation."

Post has attachment
Jeremy Howard interviewed on Lateline, a TV program in Australia. He is asked, what can AI not do? Caring for people, providing motivation, playing sports. But most things people do on most jobs can be done by machines. Self-driving cars and trucks will be good for congestion and pollution. He is not an expert in medicine, but his AI company provides tools for doctors, and there is a 10x shortage of doctors in the developing world and people are dying unnecessarily. So his idea was to use the same tools that Facebook uses to decide what ads to show you or Google uses to organize your photos to help doctors decide what treatments to give you and how to organize your diagnostic materials, and amazingly enough it turned out to work really well. The reaction from doctors has been fantastically positive. People warned him the medical industry would be conservative and try to get in his way, but he's found the opposite. Doctors get into medicine because they want to help people, and when they see what the technology can do and find they can help more people more accurately, they have a great reaction. Asked further about jobs, he says 80% of jobs are in the service sector and are ripe for automation. Computers haven't been able to do jobs like food preparation because they couldn't see. The final end state could be really great situation where we all get to spend our time doing what we want, but getting from here to there could be disruptive. The way the economy and society has to work will have to change. Asked whether machines would become smarter than humans, he said he didn't know. He looks at how technology will impact the economy based on what it can do now.

Post has attachment
"In the talk, Naftali Tishby, a computer scientist and neuroscientist from the Hebrew University of Jerusalem, presented evidence in support of a new theory explaining how deep learning works. Tishby argues that deep neural networks learn according to a procedure called the 'information bottleneck,' which he and two collaborators first described in purely theoretical terms in 1999. The idea is that a network rids noisy input data of extraneous details as if by squeezing the information through a bottleneck, retaining only the features most relevant to general concepts. Striking new computer experiments by Tishby and his student Ravid Shwartz-Ziv reveal how this squeezing procedure happens during deep learning, at least in the cases they studied. Tishby's findings have the AI community buzzing."

Post has attachment
"Lidar-equipped autonomous wheelchairs roll out in Singapore and Japan." The first is the Singapore-MIT Alliance for Research and Technology, or SMART wheelchair. "The robot's computer uses data from three lidars to make a map. A localization algorithm then determines where the smart chair is on the map. The chair's six wheels lend stability, and the chair is designed to make tight turns and fit through normal-size doorframes."

"A second autonomous wheelchair recently premiered at Haneda Airport in Tokyo, designed by Panasonic and Whill, creator of the Model A Whill wheelchair, a sleek, high-tech wheelchair now on the market in Japan and the United States."

Post has attachment
"A framework for identifying key patterns that precede an extreme event" has been devised. "Many extreme events -- from a rogue wave that rises up from calm waters, to an instability inside a gas turbine, to the sudden extinction of a previously hardy wildlife species -- seem to occur without warning."

"We have applied this framework to turbulent fluid flows, which are the Holy Grail of extreme events. They're encountered in climate dynamics in the form of extreme rainfall, in engineering fluid flows such as stresses around an airfoil, and acoustic instabilities inside gas turbines."

The researchers developed "a general framework, in the form of a computer algorithm, that combines both equations and available data to identify the precursors of extreme events that are most likely to occur in the real world." "In this way, the framework acts as a sort of sieve, capturing only those precursors that one would actually see in a real-world system."

Post has attachment
Derek Muller aka Mr. Veritasium gets into an Uber and asks the driver if he's afraid he'll lose his job to self-driving cars.

Post has shared content
Jellyfish sleep, even though they have no brains, or even spines, just neurons. "In order to be considered 'sleeping,' an organism must meet three critical criteria. First, it must demonstrate a period of reduced activity, or quiescence. Second, the organism must exhibit a decreased response to otherwise-arousing stimuli while in the quiescent state. Finally, the organism must show an increased sleep drive when it is deprived of sleep."

"This finding that jellyfish sleep implies that sleep is an ancient behavior, largely untouched by millennia of evolution."

The Washington Post article +Yonatan Zunger linked to is paywalled for me, so I'm providing this link to CalTech's article on the same research in case you have the same problem.
Sleep is one of the great mysteries of biology. It's obviously dangerous to do it – you're unaware of both predators and potential food going by – and yet it's been observed in everything down to roundworms, and without it, we die. Clearly it does something extremely important.

A discovery by three Caltech grad students adds another layer to this: Cassiopea jellyfish, also known as "upside-down jellyfish," sleep – even though they don't have brains at all, just a diffuse network of neurons. They show the same symptoms that more complex animals do: reversible quiescence (they can be woken up), increased arousal threshold, (they don't jump at things they would normally notice), and homeostatic regulation of sleep (they have trouble if they don't do it regularly).

Apparently, whatever sleep does, it's so basic to survival that even diffuse neural networks need to do it.

Post has attachment
A machine-learning algorithm to discern structural changes in the brain caused by Alzheimer's disease has been developed that purportedly can detect the disease almost a decade before doctors can diagnose the disease from symptoms alone.

"The researchers divided each brain scan into small regions and analysed the neuronal connectivity between them, without making any assumptions about the ideal size of these regions for diagnosis."

The article doesn't describe any further how the classification algorithm works, but claims in a test of 148 scan, 48 of which had Alzheimer's and 48 of which had mild cognitive impairment (MCI) but were known to have developed Alzheimer's disease 2.5 to nine years later, and it diagnosed Alzheimer's with 86% accuracy and MCI with 84% accuracy.
Wait while more posts are being loaded