Michigan Researchers Working on Smart Dust Prototypes, Dubbed “Micro Motes”
The next generation of computers will be able to carry out complex calculations but will be little bigger than a snowflake.
Such tiny computers – nicknamed smart dust – would work much like their larger cousins, says Prabal Dutta at the University of Michigan in Ann Arbor. They will have tiny CPUs that run programs on a skeleton operating system and be able to access equally small banks of RAM and flash memory.
The plan is for such sensor-packed machines to be embedded in buildings and objects in their hundreds or even thousands, providing constant updates on the world around us.
Dutta’s group is creating the first prototypes, which they have dubbed Michigan Micro Motes. These devices, a cubic millimetre in size, come equipped with sensors to monitor temperature or movement, say, and can send data via radio waves.
…Like microscopic Robinson Crusoes, the motes will live off the power they can scavenge from their surroundings. A mote near a light source might use a tiny solar panel, while a mote running somewhere with greater temperature extremes can be built to tap into that, by converting the heat energy that flows between hot and cold into electricity.
So what will be smart dust’s killer app? The Michigan team says Micro Motes could be used to monitor every tiny movement of large structures like bridges or skyscrapers. And motes in a smart house could report back on lighting, temperature, carbon monoxide levels and occupancy. With motes embedded in all of your belongings it might be possible to run a Google search in the physical world. For example, asking Google “where are my keys?” would give you the right answer if they have been fitted with a mote.
(via Smart dust computers are no bigger than a snowflake - tech - 26 April 2013 - New Scientist)
Biochemists Create Enzyme-Based Memory Capable of Learning
Electronic processors are highly efficient at certain types of computation. For example, a standard PC can vastly outperform any human at arithmetic. However, computer scientists have long been fascinated by the ability of biological systems to do tasks, such as face recognition, at speeds and a power efficiency that put the most powerful supercomputers to shame.
Clearly, biology is able of computing in ways that traditional processors have failed to capture, which is why there is a significant interest in unconventional methods of computing that explore new ways of processing information.
One form of unconventional computing is biochemical and involves using molecules to encode information and using chemical reactions to process it. Nature has developed highly complex machinery for doing this so much of the focus has been on exploiting biological molecules for this task, using proteins, DNA and the like.
Today, Vera Bocharova and a few pals at Clarkson University in Potsdam, New York, say they ‘ve used a set of enzymes to create a memory system that can “learn” to produce a specific output given a certain input. They says this system can even “unlearn” again later. “We report the first realization of a simple variant of associative memory in an enzymatic biochemical process,” they say.
(via First Enzyme-Based Memory Created in the Lab | MIT Technology Review)
Retro Computing: Inside SAGE - IBM’s 20 Acre Cold War Supercomputer
In 1957, IBM began the construction of the Semi-Automatic Ground Environment, by far the world’s largest computer. Spanning more than 20 different locations, each equipped with acre-sized computers and connected by a nation-wide network of bleeding-edge 1,300-baud modems, SAGE was the pinnacle of the United States’ Cold War radar and missile air defenses.
SAGE, like most supercomputers, was built to solve a big data problem. During the Cold War, hundreds of radar installations across North America were constantly on the lookout for Soviet missiles and bombers. As you can imagine, these stations produced a lot of data — a lot of data that needed to be analyzed and acted upon immediately.
With the physical size of the US, the high speed of modern jet aircraft, and the sheer number of possible attack vectors, the US military decided that a network of computers was the only viable solution.
SAGE consisted of 20 or so Direction Centers, each of which was a windowless, one-acre-large concrete cube. Inside each DC were two CPUs, each one measuring 7,500 sq ft and consisting of 60,000 vacuum tubes, 175,000 diodes, 13,000 newfangled transistors, and 256KB of magnetic core RAM, consuming a total of 3MW of power and weighing in at 250 tons. Each CPU — only one operated at a time; the other was kept as a hot spare to minimize downtime — was capable of executing 75,000 instructions per second, which was enough to spit out tons of radar data to 150 CRT consoles.
(via Inside IBM’s $67 billion SAGE, the largest computer ever built | ExtremeTech)
Turing Machine Made of Artificial Muscles Paves Way for Smart Prosthetics and Soft Robots
In the hierarchy of computing hardware, artificial muscle doesn’t really even register: it’s usually a target for action, not the perpetrator.
The University of Auckland has figured out a way to let those muscles play a more active role. Its prototype Turing machine uses a set of electroactive polymer muscles to push memory elements into place and squeeze piezoresistive switches, performing virtually any calculation through flexing.
The proof-of-concept computer won’t give silicon circuits any threat when it’s running at just 0.15Hz and takes up as much space as a mini fridge, but the hope is to dramatically speed up and shrink down future iterations to where there are advanced computers that occupy the same size as real muscles.
Researchers ultimately envision smart prosthetic limbs with near-natural reflexes, completely soft robots with complex gestures and even a switch from digital to analog computing for some tasks. Although we’re quite a distance away from any of those muscle-bound ideas becoming everyday realities, it’s good to at least see them on the horizon.
(via Turing machine built from artificial muscles may lead to smart prosthetics)
Australian Computer Scientists Develop Digital Face to Add Emotion to Synthesized Speech
How it works is, the user types a phrase for Zoe to say. Six sliders allow you to set the emotions; for example, you could combine happiness and anger, setting them to halfway or full strength, depending on what you want her to convey. Then you can slow down or speed up her speech, giving a pretty large array of tone.
When tested with a group of 20 volunteers, they were able to accurately guess the emotion 77 per cent of the time — more than with the real-life Zoë, with whom the success rate was 73 per cent.
The team sees Zoe being used in the future as a personal assistant, but there are other potential applications as well, because the framework for the face is very light — tens of MBs — which means that it can be incorporated into small devices. It could also enable people to upload their own faces and voices into the program; the team envisions these being used as sort of “face messages” rather than text messages.
(via Virtual talking head can express human emotions - Crave)
Genetic Computing: Researchers Split Viral Gene to Create Biological AND Gate
In recent years, researchers in the messy world of biology have been able to build systems that function like the clean, binary switches on computer chips…
Unfortunately, most of these share a significant limitation: they rely on proteins from bacteria that act as switches to turn genes on and off under specific conditions. We know about only a limited number of these genetic switches, which may set a severe limit on the number of logical operations we can string together inside a cell.
A paper in this week’s PNAS describes a system that may allow us to get around this limitation. The new method takes a protein from a virus that infects bacteria and cuts it in two, making a pair of genes (A and B) that each produce part of the mature protein. The two parts then act as a biological version of an AND logic gate, with output (in the form of protein activity) present only when both A and B interact.
When either or both A and B are missing, the output is off. In biological terms, the inputs usually involve a simple molecule that can be sensed by proteins inside a bacteria. This paper, for example, used two kinds of sugars (arabinose and lactose).
When the sugars are present, they attach to proteins inside the cell, activating genes that are controlled by those proteins. To make an AND gate, you need to design a bit of biology that can respond to both of these signals—it should be active only when both a gene regulated by arabinose and a gene regulated by lactose are each active.
(via Wetware advances: Biological logic gate built by splitting viral gene | Ars Technica)
Scientists Perform Logic Functions Using DNA Within Living Cells
By modifying a genetic toggle switch, synthetic biologists at MIT have found a way to perform logic functions inside of living cells.
Based on plasmids, circular strings of DNA, scientists devised and inserted 16 different DNA strings into Escherichia coli cells, one for each of the binary logic functions allowable in computation.
“The key to the system is the use of recombinase enzymes, which cut and rearrange promoter and terminator DNA sequences to turn them on or off. In other words, recombinase enzymes are the inputs that determine whether the output gene is transcribed.”
(via Scientists Turn Cells into Living Computers | IdeaFeed | Big Think)
Anonymous PWNS The Fed
The Federal Reserve said on Tuesday that one of its internal websites had been briefly breached by hackers, though no critical functions of the U.S. central bank were affected by the intrusion.
The admission, which raises questions about cyber security at the Fed, follows a claim that hackers linked to the activist group Anonymous had struck the Fed on Sunday, accessing personal information of more than 4,000 U.S. bank executives, which it published on the Web.
“The Federal Reserve system is aware that information was obtained by exploiting a temporary vulnerability in a website vendor product,” a Fed spokeswoman said.
“Exposure was fixed shortly after discovery and is no longer an issue. This incident did not affect critical operations of the Federal Reserve system,” the spokeswoman said, adding that all individuals effected by the breach had been contacted.
(via Fed Confirms That Hackers Breached Its Internal Site - Business Insider)
Meet Spaun: Canada’s Advanced Brain Simulator
Spaun’s brain consists of 2.5 million neurons that are broken down into a bunch of simulated cranial subsystems, including the prefrontal cortex, basal ganglia, and thalamus, which are wired together with simulated neurons that very accurately mimic the wiring of a real human brain.
The basic idea is that these subsystems behave very similarly to a real brain: Visual input is processed by the thalamus, the data is stored in the neurons, and then the basal ganglia fires off a task to a part of the cortex that’s designed to handle that task. All of this computation is performed in a physiologically accurate way, with simulated voltage spikes and neurotransmitters. Even the limitations of the human brain are simulated… with Spaun struggling to store more than a few numbers in its short-term memory.
The end result is a brain that is mechanistically simple… but which is surprisingly flexible. By implementing just a handful of very basic tasks, it’s interesting to see how complex behavior begins to emerge. There are some tantalizing hints as to how the brain evolved: starting with simple tasks, and then building upon and weaving them together to build complex functionality. In the video below, Spaun recognizes the pattern of a number sequence — the kind of question you would find on an actual IQ test.
(via Spaun, the most realistic artificial human brain yet | ExtremeTech)
Computer That Stores And Processes Information At The Same Time Is A Key Milestone on the Road to Brain Emulation
At the heart of this new form of computing are the memristor, memcapacitor and meminductor, fundamental electronic components that store information while operating as resistors, capacitors and inductors respectively.
These devices were predicted theoretically in the 1970s but only manufactured for the first time in 2008, so they are new kids on the electronics block.
Physicists immediately recognised the ability of so-called memelements to mimic biological computing and various groups have designed and built chips to exploit this idea…
But the properties of memelements that make them so good at biological computing has been hard to pin down. Which is where Di Ventra and Pershin come in. These guys have distilled the essential properties that ought to allow memelements to match the brain’s performance. They say these properties include the ability to store information over long periods; the ability to act collectively so that the state of a memdevice as a whole depends on the states of all its memelements; a robustness against noise and small imperfections; and so on.
(via The Computer That Stores and Processes Information At the Same Time | MIT Technology Review)
