Have we hit a major artificial intelligence milestone?

By Hugo Angel,

Image: REUTERS/China DailyStudents play the board game “Go”, known as “Weiqi” in Chinese, during a competition.
Google’s computer program AlphaGo has defeated a top-ranked Go player in the first round of five historic matches – marking a significant achievement in the development of artificial intelligence.
AlphaGo’s victory over a human champion shows an artificial intelligence system has mastered the most complex game ever designed. The ancient Chinese board game is vastly more complicated than chess and is said to have more possible configurations than there are atoms in the Universe.
The battle between AlphaGo, developed by Google’s Deepmind unit, and South Korea’s Lee Se-dol was said by commentators to be close, with both sides making some mistakes.
Game playing is an important way to measure AI advances, demonstrating that machines can outperform humans at intellectual tasks.
AlphaGo’s win follows in the footsteps of the legendary 1997 victory of IBM supercomputer Deep Blue over world chess champion Garry Kasparov. But Go, which relies heavily on players’ intuition to choose among vast numbers of board positions, is far more challenging for artificial intelligence than chess.
Speaking in the lead-up to the first match, Se-dol, who is currently ranked second in the world behind fellow South Korean Lee Chang-ho, said: “Having learned today how its algorithms narrow down possible choices, I have a feeling that AlphaGo can imitate human intuition to a certain degree.”
Demis Hassabis, founder and CEO of DeepMind, which was acquired by Google in 2014, previously described “Go as the pinnacle of game AI research” and the “holy grail” of AI since Deep Blue beat Kasparov.

Experts had predicted it would take another decade for AI systems to beat professional Go players. But in January, the journal Nature reported that AlphaGo won a five-game match against European champion Fan Hui. Since then the computer program’s performance has steadily improved.
Mastering the game of Go. Nature
While DeepMind’s team built AlphaGo to learn in a more human-like way, it still needs much more practice than a human expert, millions of games rather than thousands.
Potential future uses of AI programs like AlphaGo could include improving smartphone assistants such as Apple’s Siri, medical diagnostics, and possibly even working with human scientists in research.
by Rosamond Hutt, Senior Producer, Formative Content
9 March 2016

Monster Machine Cracks The Game Of Go

By Hugo Angel,

Illustration: Google DeepMind/Nature
A computer program has defeated a master of the ancient Chinese game of Go, achieving one of the loftiest of the Grand Challenges of AI at least a decade earlier than anyone had thought possible.
The programmers, at Google’s Deep Mind laboratory, in London, .write in today’s issue of .Nature that their program .AlphaGo defeated .Fan Hui, the European Go champion, 5 games to nil, in a match held last October in the company’s offices. Earlier, the program had won 494 out of 495 games against the best rival Go programs.
AlphaGo’s creators now hope to seal their victory at a 5-game match against .Lee Se-dol, the best Go player in the world. That match, for a $1 million prize fund, is scheduled to take place in March in Seoul, South Korea.
The program’s victory marks the rise not merely of the machines but of new methods of computer programming based on self-training neural networks. In support of their claim that this method can be applied broadly, the researchers cited their success, which we .reported a year ago, in getting neural networks to learn how to play an entire set of video games from Atari. Future applications, they say, may include financial trading, climate modeling and medical diagnosis.
Not all of AlphaGo’s skill is self-taught. First, the programmers jumpstarted the training by having the program predict moves in a database of master games. It eventually reached a success rate of 57 percent, versus 44 percent for the best rival programs.
Then, to go beyond mere human performance, the program conducted its own research through a trial-and-error approach that involved playing millions of games against itself. In this fashion it discovered, one by one, many of the rules of thumb that textbooks have been imparting to Go students for centuries. Google DeepMind calls the self-guided method reinforced learning, but it’s really just another word for “.deep learning,” the current AI buzzword.
Not only can self-trained machines surpass the game-playing powers of their creators, they can do so in ways that programmers can’t even explain. It’s a different world from the one that AI’s founders envisaged decades ago.
Commenting on the death yesterday of AI pioneer Marvin Minksy, Demis Hassabis, the lead author of the Nature paper, said “It would be interesting to see what he would have said,” said Hassabis. “I suspect he would have been pretty surprised at how quickly this has arrived.
That’s because, as programmers would say, Go is such a bear. Then again, even chess was a bear, at first. Back in 1957, the late Herbert Simon famously predicted that a computer would beat the world champion at chess within a decade. But it was only in 1997 that World Chess Champion Garry Kasparov lost to IBM’s Deep Blue—a multimillion-dollar, purpose-built machine that filled a small room. Today you can .download a $100 program to a decently powered laptop and watch it utterly cream any chess player in the world.
Go is harder for machines because the positions are harder to judge and there are a whole lot more positions.
Judgement is harder because the pieces, or “stones,” are all of equal value, whereas those in chess have varying values—a Queen, for instance, is worth nine times more than a pawn, on average. Chess programmers can thus add up those values (and throw in numerical estimates for the placement of pieces and pawns) to arrive at a quick-and-dirty score of a game position. No such expedient exists for Go.
There are vastly more positions to judge than in chess because Go offers on average 10 times more options at every move and there are about three times as many moves in an game. The number of possible board configurations in Go is estimated at 10 to the 170th power—“more than the number of atoms in the universe,” said Hassabis.
Some researchers .tried to adapt to Go some of the forward-search techniques devised for chess; others .relied on random simulations of games in the aptly named Monte Carlo method. The Google DeepMind people leapfrogged them all with deep, or convolutional, neural networks, so named because they imitate the brain (up to a point).
A neural network links units that are the computing equivalent to a brain’s neurons—first by putting them into layers, then by stacking the layers. AlphaGo’s are 12 layers deep. Each “neuron” connects to its neighbors in its own layer and also those in the layers directly above and below it. A signal sent to one neuron causes it to strengthen or weaken its connections to other ones, so over time, the network changes its configuration.
To train the system

  • you first expose it to input data
  • Next, you test the output signal against the metric you’re using—say, predicting a master’s move—and 
  • reinforce correct decisions by strengthening the underlying connections. 
  • Over time, the system produces better outputs. You might say that it learns.
AlphaGo has two networks

  • The policy network cuts down on the number of moves to look at, and 
  • the evaluation network allows you to cut short the depth of that search,” or the number of moves the machine must look ahead, Hassabis said. 

 “Both neural networks together make the search tractable.”

The main difference from the system that played Atari is the inclusion of a search-ahead function: “In Atari you can do well by reacting quickly to current events,” said Hassabis. “In Go you need a plan.”
After exhaustive training the two networks, taken by themselves, could play Go as well as any program did. But when the researchers coupled the neural networks to a forward-searching algorithm, the machine was able to dominate rival programs completely. Not only did it win all but one of the hundreds of games it played against them, it was even able to give them a handicap of four extra moves, made at the beginning of the game, and still beat them.
About that one defeat: “The search has stochastic [random] element, so there’s always a possibility that it will make a mistake,” David Silver said. “As we improve, we reduce probability of making a mistake, but mistakes happen. As in that one particular game.
Anyone might cock an eyebrow at the claim that AlphaGo will have practical spin-offs. Games programmers have often justified their work by promising such things but so far they’ve had little to show for their efforts. IBM’s Deep Blue did nothing but play chess, and IBM’s Watson—.designed to beat the television game show Jeopardy!—will need laborious retraining to be of service in its .next appointed task of helping doctors diagnose and treat patients.
But AlphaGo’s creators say that because of the generalized nature of their approach, direct spin-offs really will come—this time for sure. And they’ll get started on them just as soon as the March match against the world champion is behind them.
Posted 27 Jan 2016

Brain waves may be spread by weak electrical field

By Hugo Angel,

The research team says the electrical fields could be behind the spread of sleep and theta waves, along with epileptic seizure waves (Credit:Shutterstock)
Mechanism tied to waves associated with epilepsy
Researchers at Case Western Reserve University may have found a new way information is communicated throughout the brain.
Their discovery could lead to identifying possible new targets to investigate brain waves associated with memory and epilepsy and better understand healthy physiology.
They recorded neural spikes traveling at a speed too slow for known mechanisms to circulate throughout the brain. The only explanation, the scientists say, is the wave is spread by a mild electrical field they could detect. Computer modeling and in-vitro testing support their theory.
Others have been working on such phenomena for decades, but no one has ever made these connections,” said Steven J. Schiff, director of the Center for Neural Engineering at Penn State University, who was not involved in the study. “The implications are that such directed fields can be used to modulate both pathological activities, such as seizures, and to interact with cognitive rhythms that help regulate a variety of processes in the brain.
Scientists Dominique Durand, Elmer Lincoln Lindseth Professor in Biomedical Engineering at Case School of Engineering and leader of the research, former graduate student Chen Sui and current PhD students Rajat Shivacharan and Mingming Zhang, report their findings in The Journal of Neuroscience.
Researchers have thought that the brain’s endogenous electrical fields are too weak to propagate wave transmission,” Durand said. “But it appears the brain may be using the fields to communicate without synaptic transmissions, gap junctions or diffusion.
How the fields may work
Computer modeling and testing on mouse hippocampi (the central part of the brain associated with memory and spatial navigation) in the lab indicate the field begins in one cell or group of cells.
Although the electrical field is of low amplitude, the field excites and activates immediate neighbors, which, in turn, excite and activate immediate neighbors, and so on across the brain at a rate of about 0.1 meter per second.
Blocking the endogenous electrical field in the mouse hippocampus and increasing the distance between cells in the computer model and in-vitro both slowed the speed of the wave.
These results, the researchers say, confirm that the propagation mechanism for the activity is consistent with the electrical field.
Because sleep waves and theta waves–which are associated with forming memories during sleep–and epileptic seizure waves travel at about 1 meter per second, the researchers are now investigating whether the electrical fields play a role in normal physiology and in epilepsy.
If so, they will try to discern what information the fields may be carrying. Durand’s lab is also investigating where the endogenous spikes come from.
ORIGINAL: Eurkalert

Memory capacity of brain is 10 times more than previously thought

By Hugo Angel,

Data from the Salk Institute shows brain’s memory capacity is in the petabyte range, as much as entire Web

LA JOLLA—Salk researchers and collaborators have achieved critical insight into the size of neural connections, putting the memory capacity of the brain far higher than common estimates. The new work also answers a longstanding question as to how the brain is so energy efficient and could help engineers build computers that are incredibly powerful but also conserve energy.
This is a real bombshell in the field of neuroscience,” said Terry Sejnowski from the Salk Institute for Biological Studies. “Our new measurements of the brain’s memory capacity increase conservative estimates by a factor of 10 to at least a petabyte (215 Bytes = 1000 TeraBytes), in the same ballpark as the World Wide Web.
Our memories and thoughts are the result of patterns of electrical and chemical activity in the brain. A key part of the activity happens when branches of neurons, much like electrical wire, interact at certain junctions, known as synapses. An output ‘wire’ (an axon) from one neuron connects to an input ‘wire’ (a dendrite) of a second neuron. Signals travel across the synapse as chemicals called neurotransmitters to tell the receiving neuron whether to convey an electrical signal to other neurons. Each neuron can have thousands of these synapses with thousands of other neurons.
When we first reconstructed every dendrite, axon, glial process, and synapse from a volume of hippocampus the size of a single red blood cell, we were somewhat bewildered by the complexity and diversity amongst the synapses,” says Kristen Harris, co-senior author of the work and professor of neuroscience at the University of Texas, Austin. “While I had hoped to learn fundamental principles about how the brain is organized from these detailed reconstructions, I have been truly amazed at the precision obtained in the analyses of this report.
Synapses are still a mystery, though their dysfunction can cause a range of neurological diseases. Larger synapses—with more surface area and vesicles of neurotransmitters—are stronger, making them more likely to activate their surrounding neurons than medium or small synapses.
The Salk team, while building a 3D reconstruction of rat hippocampus tissue (the memory center of the brain), noticed something unusual. In some cases, a single axon from one neuron formed two synapses reaching out to a single dendrite of a second neuron, signifying that the first neuron seemed to be sending a duplicate message to the receiving neuron.
At first, the researchers didn’t think much of this duplicity, which occurs about 10 percent of the time in the hippocampus. But Tom Bartol, a Salk staff scientist, had an idea: if they could measure the difference between two very similar synapses such as these, they might glean insight into synaptic sizes, which so far had only been classified in the field as small, medium and large.
In a computational reconstruction of brain tissue in the hippocampus, Salk scientists and UT-Austin scientists found the unusual occurrence of two synapses from the axon of one neuron (translucent black strip) forming onto two spines on the same dendrite of a second neuron (yellow). Separate terminals from one neuron’s axon are shown in synaptic contact with two spines (arrows) on the same dendrite of a second neuron in the hippocampus. The spine head volumes, synaptic contact areas (red), neck diameters (gray) and number of presynaptic vesicles (white spheres) of these two synapses are almost identical. Credit: Salk Institut
To do this, researchers used advanced microscopy and computational algorithms they had developed to image rat brains and reconstruct the connectivity, shapes, volumes and surface area of the brain tissue down to a nanomolecular level.
The scientists expected the synapses would be roughly similar in size, but were surprised to discover the synapses were nearly identical.
We were amazed to find that the difference in the sizes of the pairs of synapses were very small, on average, only about 8 percent different in size,” said Tom Bartol, one of the scientists. “No one thought it would be such a small difference. This was a curveball from nature.
Because the memory capacity of neurons is dependent upon synapse size, this eight percent difference turned out to be a key number the team could then plug into their algorithmic models of the brain to measure how much information could potentially be stored in synaptic connections.
It was known before that the range in sizes between the smallest and largest synapses was a factor of 60 and that most are small.
But armed with the knowledge that synapses of all sizes could vary in increments as little as eight percent between sizes within a factor of 60, the team determined there could be about 26 categories of sizes of synapses, rather than just a few.
Our data suggests there are 10 times more discrete sizes of synapses than previously thought,” says Bartol. In computer terms, 26 sizes of synapses correspond to about 4.7 “bits” of information. Previously, it was thought that the brain was capable of just one to two bits for short and long memory storage in the hippocampus.
This is roughly an order of magnitude of precision more than anyone has ever imagined,” said Sejnowski.
What makes this precision puzzling is that hippocampal synapses are notoriously unreliable. When a signal travels from one neuron to another, it typically activates that second neuron only 10 to 20 percent of the time.
We had often wondered how the remarkable precision of the brain can come out of such unreliable synapses,” says Bartol. One answer, it seems, is in the constant adjustment of synapses, averaging out their success and failure rates over time. The team used their new data and a statistical model to find out how many signals it would take a pair of synapses to get to that eight percent difference.
The researchers calculated that
  • for the smallest synapses, about 1,500 events cause a change in their size/ability (20 minutes) and
  • for the largest synapses, only a couple hundred signaling events (1 to 2 minutes) cause a change.
This means that every 2 or 20 minutes, your synapses are going up or down to the next size,” said Bartol. “The synapses are adjusting themselves according to the signals they receive.
From left: Terry Sejnowski, Cailey Bromer and Tom Bartol. Credit: Salk Institute
Our prior work had hinted at the possibility that spines and axons that synapse together would be similar in size, but the reality of the precision is truly remarkable and lays the foundation for whole new ways to think about brains and computers,” says Harris. “The work resulting from this collaboration has opened a new chapter in the search for learning and memory mechanisms.” Harris adds that the findings suggest more questions to explore, for example, if similar rules apply for synapses in other regions of the brain and how those rules differ during development and as synapses change during the initial stages of learning.
The implications of what we found are far-reaching. Hidden under the apparent chaos and messiness of the brain is an underlying precision to the size and shapes of synapses that was hidden from us.
The findings also offer a valuable explanation for the brain’s surprising efficiency. The waking adult brain generates only about 20 watts of continuous power—as much as a very dim light bulb. The Salk discovery could help computer scientists build ultra-precise but energy-efficient computers, particularly ones that employ deep learning and neural nets techniques capable of sophisticated learning and analysis, such as speech, object recognition and translation.
This trick of the brain absolutely points to a way to design better computers,”said Sejnowski. “Using probabilistic transmission turns out to be as accurate and require much less energy for both computers and brains.
Other authors on the paper were Cailey Bromer of the Salk Institute; Justin Kinney of the McGovern Institute for Brain Research; and Michael A. Chirillo and Jennifer N. Bourne of the University of Texas, Austin.
The work was supported by the NIH and the Howard Hughes Medical Institute.
ORIGINAL: Salk.edu
January 20, 2016

Bridging the Bio-Electronic Divide

By Hugo Angel,

New effort aims for fully implantable devices able to connect with up to one million neurons
A new DARPA program aims to develop an implantable neural interface able to provide unprecedented signal resolution and data-transfer bandwidth between the human brain and the digital world. The interface would serve as a translator, converting between the electrochemical language used by neurons in the brain and the ones and zeros that constitute the language of information technology. The goal is to achieve this communications link in a biocompatible device no larger than one cubic centimeter in size, roughly the volume of two nickels stacked back to back.
The program, Neural Engineering System Design (NESD), stands to dramatically enhance research capabilities in neurotechnology and provide a foundation for new therapies.
“Today’s best brain-computer interface systems are like two supercomputers trying to talk to each other using an old 300-baud modem,” said Phillip Alvelda, the NESD program manager. “Imagine what will become possible when we upgrade our tools to really open the channel between the human brain and modern electronics.
Among the program’s potential applications are devices that could compensate for deficits in sight or hearing by feeding digital auditory or visual information into the brain at a resolution and experiential quality far higher than is possible with current technology.
Neural interfaces currently approved for human use squeeze a tremendous amount of information through just 100 channels, with each channel aggregating signals from tens of thousands of neurons at a time. The result is noisy and imprecise. In contrast, the NESD program aims to develop systems that can communicate clearly and individually with any of up to one million neurons in a given region of the brain.
Achieving the program’s ambitious goals and ensuring that the envisioned devices will have the potential to be practical outside of a research setting will require integrated breakthroughs across numerous disciplines including 
  • neuroscience, 
  • synthetic biology, 
  • low-power electronics, 
  • photonics, 
  • medical device packaging and manufacturing, systems engineering, and 
  • clinical testing.
In addition to the program’s hardware challenges, NESD researchers will be required to develop advanced mathematical and neuro-computation techniques to first transcode high-definition sensory information between electronic and cortical neuron representations and then compress and represent those data with minimal loss of fidelity and functionality.
To accelerate that integrative process, the NESD program aims to recruit a diverse roster of leading industry stakeholders willing to offer state-of-the-art prototyping and manufacturing services and intellectual property to NESD researchers on a pre-competitive basis. In later phases of the program, these partners could help transition the resulting technologies into research and commercial application spaces.
To familiarize potential participants with the technical objectives of NESD, DARPA will host a Proposers Day meeting that runs Tuesday and Wednesday, February 2-3, 2016, in Arlington, Va. The Special Notice announcing the Proposers Day meeting is available at https://www.fbo.gov/spg/ODA/DARPA/CMO/DARPA-SN-16-16/listing.html. More details about the Industry Group that will support NESD is available at https://www.fbo.gov/spg/ODA/DARPA/CMO/DARPA-SN-16-17/listing.html. A Broad Agency Announcement describing the specific capabilities sought will be forthcoming on www.fbo.gov.
NESD is part of a broader portfolio of programs within DARPA that support President Obama’s brain initiative. For more information about DARPA’s work in that domain, please visit:http://www.darpa.mil/program/our-research/darpa-and-the-brain-initiative.
[email protected]

Microsoft Neural Net Shows Deep Learning can get Way Deeper

By Hugo Angel,

Silicon Wafer by Sonic
COMPUTER VISION IS now a part of everyday life. Facebook recognizes faces in the photos you post to the popular social network. The Google Photos app can find images buried in your collection, identifying everything from dogs to birthday parties to gravestones. Twitter can pinpoint pornographic images without help from human curators.
All of this “seeing” stems from a remarkably effective breed of artificial intelligence called deep learning. But as far as this much-hyped technology has come in recent years, a new experiment from Microsoft Research shows it’s only getting started. Deep learning can go so much deeper.
We’re staring at a huge design space, trying to figure out where to go next.‘ 


This revolution in computer vision was a long time coming. A key turning point came in 2012, when artificial intelligence researchers from the University of Toronto won a competition called ImageNet. ImageNet pits machines against each other in an image recognition contest—which computer can identify cats or cars or clouds more accurately?—and that year, the Toronto team, including researcher Alex Krizhevsky and professor Geoff Hinton, topped the contest using deep neural nets, a technology that learns to identify images by examining enormous numbers of them, rather than identifying images according to rules diligently hand-coded by humans.
Toronto’s win provided a roadmap for the future of deep learning. In the years since, the biggest names on the ‘net—including Facebook, Google, Twitter, and Microsoft—have used similar tech to build computer vision systems that can match and even surpass humans. “We can’t claim that our system ‘sees’ like a person does,” says Peter Lee, the head of research at Microsoft. “But what we can say is that for very specific, narrowly defined tasks, we can learn to be as good as humans.
Roughly speaking, neural nets use hardware and software to approximate the web of neurons in the human brain. This idea dates to the 1980s, but in 2012, Krizhevsky and Hinton advanced the technology by running their neural nets atop graphics processing units, or GPUs. These specialized chips were originally designed to render images for games and other highly graphical software, but as it turns out, they’re also suited to the kind of math that drives neural nets. Google, Facebook, Twitter, Microsoft, and so many others now use GPU-powered-AI to handle image recognition and so many others tasks, from Internet search to security. Krizhevsky and Hinton joined the staff at Google.
Deep learning can go so much deeper.
Now, the latest ImageNet winner is pointing to what could be another step in the evolution of computer vision—and the wider field of artificial intelligence. Last month, a team of Microsoft researchers took the ImageNet crown using a new approach they call a deep residual network. The name doesn’t quite describe it. They’ve designed a neural net that’s significantly more complex than typical designs—one that spans 152 layers of mathematical operations, compared to the typical six or seven. It shows that, in the years to come, companies like Microsoft will be able to use vast clusters of GPUs and other specialized chips to significantly improve not only image recognition but other AI services, including systems that recognize speech and even understand language as we humans naturally speak it.
In other words, deep learning is nowhere close to reaching its potential. “We’re staring at a huge design space,” Lee says, “trying to figure out where to go next.
Layers of Neurons
Deep neural networks are arranged in layers. Each layer is a different set of mathematical operations—aka algorithms. The output of one layer becomes the input of the next. Loosely speaking, if a neural network is designed for image recognition, one layer will look for a particular set of features in an image—edges or angles or shapes or textures or the like—and the next will look for another set. These layers are what make these neural networks deep. “Generally speaking, if you make these networks deeper, it becomes easier for them to learn,” says Alex Berg, a researcher at the University of North Carolina who helps oversee the ImageNet competition.
Constructing this kind of mega-neural net is flat-out difficult.
Today, a typical neural network includes six or seven layers. Some might extend to 20 or even 30. But the Microsoft team, led by researcher Jian Sun, just expanded that to 152. In essence, this neural net is better at recognizing images because it can examine more features. “There is a lot more subtlety that can be learned,” Lee says.
In the past, according Lee and researchers outside of Microsoft, this sort of very deep neural net wasn’t feasible. Part of the problem was that as your mathematical signal moved from layer to layer, it became diluted and tended to fade. As Lee explains, Microsoft solved this problem by building a neural net that skips certain layers when it doesn’t need them, but uses them when it does. “When you do this kind of skipping, you’re able to preserve the strength of the signal much further,” Lee says, “and this is turning out to have a tremendous, beneficial impact on accuracy.
Berg says that this is an notable departure from previous systems, and he believes that others companies and researchers will follow suit.
Deep Difficulty
The other issue is that constructing this kind of mega-neural net is tremendously difficult. Landing on a particular set of algorithms—determining how each layer should operate and how it should talk to the next layer—is an almost epic task. But Microsoft has a trick here, too. It has designed a computing system that can help build these networks.
As Jian Sun explains it, researchers can identify a promising arrangement for massive neural networks, and then the system can cycle through a range of similar possibilities until it settles on this best one. “In most cases, after a number of tries, the researchers learn [something], reflect, and make a new decision on the next try,” he says. “You can view this as ‘human-assisted search.’”
Microsoft has designed a computing system that can help build these networks.
According to Adam Gibson—the chief researcher at deep learning startup Skymind—this kind of thing is getting more common. It’s called “hyper parameter optimization.” “People can just spin up a cluster [of machines], run 10 models at once, find out which one works best and use that,” Gibson says. “They can input some baseline parameter—based on intuition—and the machines kind of homes in on what the best solution is.” As Gibson notes, last year Twitter acquired a company, Whetlab, that offers similar ways of “optimizing” neural networks.

‘A Hardware Problem’
As Peter Lee and Jian Sun describe it, such an approach isn’t exactly “brute forcing” the problem. “With very very large amounts of compute resources, one could fantasize about a gigantic ‘natural selection’ setup where evolutionary forces help direct a brute-force search through a huge space of possibilities,” Lee says. “The world doesn’t have those computing resources available for such a thing…For now, we will still depend on really smart researchers like Jian.
But Lee does say that, thanks to new techniques and computer data centers filled with GPU machines, the realm of possibilities for deep learning are enormous. A big part of the company’s task is just finding the time and the computing power needed to explore these possibilities. “This work as dramatically exploded the design space. The amount of ground to cover, in terms of scientific investigation, has become exponentially larger,” Lee says. And this extends well beyond image recognition, into speech recognition, natural language understanding, and other tasks.
As Lee explains, that’s one reason Microsoft is not only pushing to improve the power of its GPUs clusters, but exploring the use of other specialized processors, including FPGAs—chips that can programmed for particular tasks, such as deep learning. “There has also been an explosion in demand for much more experimental hardware platforms from our researchers,” he says. And this work is sending ripples across the wider of world of tech and artificial intelligence. This past summer, in its largest ever acquisition deal, Intel agreed to buy Altera, which specializes in FPGAs.
Indeed, Gibson says that deep learning has become more of “a hardware problem.” Yes, we still need top researchers to guide the creation of neural networks, but more and more, finding new paths is a matter of brute-forcing new algorithms across ever more powerful collections of hardware. As Gibson point out, though these deep neural nets work extremely well, we don’t quite know why they work. The trick lies in finding the complex combination of algorithms that work the best. More and better hardware can shorten the path.
The end result is that the companies that can build the most powerful networks of hardware are the companies will come out ahead. That would be Google and Facebook and Microsoft. Those that are good at deep learning today will only get better.

NVIDIA DRIVE PX 2. NVIDIA Accelerates Race to Autonomous Driving at CES 2016

By Hugo Angel,

NVIDIA today shifted its autonomous-driving leadership into high gear.
At a press event kicking off CES 2016, we unveiled artificial-intelligence technology that will let cars sense the world around them and pilot a safe route forward.
Dressed in his trademark black leather jacket, speaking to a crowd of some 400 automakers, media and analysts, NVIDIA CEO Jen-Hsun Huang revealed DRIVE PX 2, an automotive supercomputing platform that processes 24 trillion deep learning operations a second. That’s 10 times the performance of the first-generation DRIVE PX, now being used by more than 50 companies in the automotive world.
The new DRIVE PX 2 delivers 8 teraflops of processing power. It has the processing power of 150 MacBook Pros. And it’s the size of a lunchbox in contrast to earlier autonomous-driving technology being used today, which takes up the entire trunk of a mid-sized sedan.
Self-driving cars will revolutionize society,” Huang said at the beginning of his talk. “And NVIDIA’s vision is to enable them.
Volvo to Deploy DRIVE PX in Self-Driving SUVs
As part of its quest to eliminate traffic fatalities, Volvo will be the first automaker to deploy DRIVE PX 2.
Huang announced that Volvo – known worldwide for safety and reliability – will be the first automaker to deploy DRIVE PX 2.
In the world’s first public trial of autonomous driving, the Swedish automaker next year will lease 100 XC90 luxury SUVs outfitted with DRIVE PX 2 technology. The technology will help the vehicles drive autonomously around Volvo’s hometown of Gothenburg, and semi-autonomously elsewhere.
DRIVE PX 2 has the power to harness a host of sensors to get a 360 degree view of the environment around the car.
The rear-view mirror is history,” Jen-Hsun said.
Drive Safely, by Not Driving at All
Not so long ago, pundits had questioned the safety of technology in cars. Now, with Volvo incorporating autonomous vehicles into its plan to end traffic fatalities, that script has been flipped. Autonomous cars may be vastly safer than human-piloted vehicles.
Car crashes – an estimated 93 percent of them caused by human error kill 1.3 million drivers each year. More American teenagers die from texting while driving than any other cause, including drunk driving.
There’s also a productivity issue. Americans waste some 5.5 billion hours of time each year in traffic, costing the U.S. about $121 billion, according to an Urban Mobility Report from Texas A&M. And inefficient use of roads by cars wastes even vaster sums spent on infrastructure.
Deep Learning Hits the Road
Self-driving solutions based on computer vision can provide some answers. But tackling the infinite permutations that a driver needs to react to – stray pets, swerving cars, slashing rain, steady road construction crews – is far too complex a programming challenge.
Deep learning enabled by NVIDIA technology can address these challenges. A highly trained deep neural network – residing on supercomputers in the cloud – captures the experience of many tens of thousands of hours of road time.
Huang noted that a number of automotive companies are already using NVIDIA’s deep learning technology to power their efforts, getting speedup of 30-40X in training their networks compared with other technology. BMW, Daimler and Ford are among them, along with innovative Japanese startups like Preferred Networks and ZMP. And Audi said it was able in four hours to do training that took it two years with a competing solution.
  NVIDIA DRIVE PX 2 is part of an end-to-end platform that brings deep learning to the road.
NVIDIA’s end-to-end solution for deep learning starts with NVIDIA DIGITS, a supercomputer that can be used to train digital neural networks by exposing them to data collected during that time on the road. On the other end is DRIVE PX 2, which draws on this training to make inferences to enable the car to progress safely down the road. In the middle is NVIDIA DriveWorks, a suite of software tools, libraries and modules that accelerates development and testing of autonomous vehicles.
DriveWorks enables sensor calibration, acquisition of surround data, synchronization, recording and then processing streams of sensor data through a complex pipeline of algorithms running on all of the DRIVE PX 2’s specialized and general-purpose processors.
During the event, Huang reminded the audience that machines are already beating humans at tasks once considered impossible for computers, such as image recognition. Systems trained with deep learning can now correctly classify images more than 96 percent of the time, exceeding what humans can do on similar tasks.
He used the event to show what deep learning can do for autonomous vehicles.
A series of demos drove this home, showing in three steps how DRIVE PX 2 harnesses a host of sensors – lidar, radar and cameras and ultrasonic – to understand the world around it, in real time, and plan a safe and efficient path forward.
The World’s Biggest Infotainment System
The highlight of the demos was what Huang called the world’s largest car infotainment system — an elegant block the size of a medium-sized bedroom wall mounted with a long horizontal screen and a long vertical one.
While a third larger screen showed the scene that a driver would take in, the wide demo screen showed how the car — using deep learning and sensor fusion — “viewed” the very same scene in real-time, stitched together from its array of sensors. On its right, the huge portrait-oriented screen shows a highly precise map that marked the car’s progress.
It’s a demo that will leave an impression on an audience that’s going to be hear a lot about the future of driving in the week ahead.
Photos from Our CES 2016 Press Event
By Bob Sherbin on January 3, 2016

Robots are learning from YouTube tutorials

By Hugo Angel,

Do it yourself, robot. (Reuters/Kim Kyung-Hoon)
For better or worse, we’ve taught robots to mimic human behavior in countless ways. They can perform tasks as rudimentary as picking up objects, or as creative as dreaming their own dreams. They can identify bullying, and even play jazz. Now, we’ve taught robots the most human task of all: how to teach themselves to make Jell-O shots from watching YouTube videos.
Ever go to YouTube and type in something like, “How to make pancakes,” or, “How to mount a TV”? Sure you have. While many such tutorials are awful—and some are just deliberately misleading—the sheer number of instructional videos offers strong odds of finding one that’s genuinely helpful. And when all those videos are aggregated and analyzed simultaneously, it’s not hard for a robot to figure out what the correct steps are.

Researchers at Cornell University have taught robots to do just that with a system called RoboWatch. By watching and scanning multiple videos of the same “how-to” activity (with subtitles enabled), bots can 

  • identify common steps, 
  • put them in order, and 
  • learn how to do whatever the tutorials are teaching.
Robot learning is not new, but what’s unusual here is that these robots can learn without human supervision, as Phys.Org points out.
Similar research usually requires human overseers to introduce and explain words, or captions, for the robots to parse. RoboWatch, however, needs no human help, save that someone ensure all the videos analyzed fall into a single category (pdf). The idea is that a human could one day tell a robot to perform a task and then the robot would independently research and learn how to carry out that task.
So next time you getting frustrated watching a video on how to change a tire, don’t fret. Soon, a robot will do all that for you. We just have to make sure it doesn’t watch any videos about “how to take over the world.
December 22, 2015

Scientists have discovered brain networks linked to intelligence for the first time

By Hugo Angel,

Neurons Shutterstock 265323554_1024
And we may even be able to manipulate them.
For the first time ever, scientists have identified clusters of genes in the brain that are believed to be linked to human intelligence.
The two clusters, called M1 and M3, are networks each consisting of hundreds of individual genes, and are thought to influence our

  • cognitive functions, including 
  • memory, 
  • attention, 
  • processing speed, and 
  • reasoning.
Most provocatively, the researchers who identified M1 and M3 say that these clusters are probably under the control of master switches that regulate how the gene networks function. If this hypothesis is correct and scientists can indeed find these switches, we might even be able to manipulate our genetic intelligence and boost our cognitive capabilities.
“We know that genetics plays a major role in intelligence but until now haven’t known which genes are relevant,said neurologist Michael Johnson, at Imperial College London in the UK. “This research highlights some of the genes involved in human intelligence, and how they interact with each other.
The researchers made their discovery by examining the brains of patients who had undergone neurosurgery for the treatment of epilepsy. They analysed thousands of genes expressed in the brain and combined the findings with two sets of data: genetic information from healthy people who had performed IQ tests, and from people with neurological disorders and intellectual disability.
Comparing the results, the researchers discovered that some of the genes that influence human intelligence in healthy people can also cause significant neurological problems if they end up mutating.
Traits such as intelligence are governed by large groups of genes working together – like a football team made up of players in different positions,said Johnson. “We used computer analysis to identify the genes in the human brain that work together to influence our cognitive ability to make new memories or sensible decisions when faced with lots of complex information. We found that some of these genes overlap with those that cause severe childhood onset epilepsy or intellectual disability.
The research, which is reported in Nature Neuroscience, is at an early stage, but the authors believe their analysis could have a significant impact – not only on how we understand and treat brain diseases, but one day perhaps altering brainpower itself.
Eventually, we hope that this sort of analysis will provide new insights into better treatments for neurodevelopmental diseases such as epilepsy, and ameliorate or treat the cognitive impairments associated with these devastating diseases,” said Johnson. “Our research suggests that it might be possible to work with these genes to modify intelligence, but that is only a theoretical possibility at the moment – we have just taken a first step along that road.
ORIGINAL: Science Alert
22 DEC 2015

Scientists have built a functional ‘hybrid’ logic gate for use in quantum computers

By Hugo Angel,

NIST Quantum Gate
An ion trap used in NIST quantum computing experiments. Credit: Blakestad/NIST
Here’s how to solve the problem of quantum memory.
As conventional computers draw ever closer to their theoretical limit, the race is on to build a machine that can truly harness the unprecedented processing power of quantum computing. And now two research teams have independently demonstrated how entangling atoms from different elements can address the problem of quantum memory errors while functioning within a logic gate framework, and also pass the all-important test of true entanglement. 
Hybrid quantum computers allow the unique advantages of different types of quantum systems to be exploited together in a single platform,said lead author Ting Rei Tan. “Each ion species is unique, and certain ones are better suited for certain tasks such as memory storage, while others are more suited to provide interconnects for data transfer between remote systems.
In the computers we use today, data is processed and stored as binary bits, with each individual bit taking on a state of either 0 or 1. Because these states are set, there’s a finite amount of information that can ultimately be processed, and we’re quickly approaching the point where this isn’t going to be enough.
Quantum computers, on the other hand, store data as qubits, which can be in the state of 0 or 1, or can take on another state called superposition, which allows them to be both 0 and 1 at the same time. If we can figure out how to build a machine that integrates this phenomenon with data-processing capabilities, we’re looking at computers that are hundreds of millions of times faster than the super computers of today.
The qubits used in this set-up are actually atomic ions (atoms with an electron removed), and their states are determined by their spin – spin up is 1, spin down is 0. Each atomic ion is paired off, and if the control ion takes on the state of superposition, it will become entangled with its partner, so anything you do to one ion will affect the other.
This can pose problems, particularly when it comes to memory, and there’s no point storing and processing information if you can’t reliably retain it. If you’ve got an entire system built on pairs of the same atomic ions, you leave yourself open to constant errors, because if one ion is affected by a malfunction, this will also affect its partner. At the same time, using the same atomic ions in a pair makes it very difficult for them to perform separate functions.
So researchers from the University of Oxford in the UK, and a second team from the National Institute of Standards and Technology (NIST) and the University of Washington, have figured out which combinations of different elements can function together as pairs in a quantum set-up.
Each trapped ion is used to represent one ‘quantum bit’ of information. The quantum states of the ions are controlled with laser pulses of precise frequency and duration,says one of the researchers, David Lucas from the University of Oxford. “Two different species of ion are needed in the computer

  • one to store information, a ‘memory qubit’, and 
  • one to link different parts of the computer together via photons, an ‘interface qubit’.
While the Oxford team achieved this using two different isotopes of calcium (the abundant isotope calcium-40 and the rare isotope calcium-43), the second team went even further by pairing up entirely different atoms – magnesium and beryllium. Each one is sensitive to a different wavelength of light, which means zapping one with a laser pulse to control its function won’t affect its partner.
The teams them went on to demonstrate for the first time that these pairs could have their 0,1, or superposition states controlled by two different types of logic gates, called the CNOT gate and the SWAP gate. Logic gates are crucial components of any digital circuit, because they’re able to record two input values and provide a new output based on programmed logic. 
A CNOT gate flips the second (target) qubit if the first (control) qubit is a 1; if it is a 0, the target bit is unchanged,the NIST press release explains. “If the control qubit is in a superposition, the ions become entangled. A SWAP gate interchanges the qubit states, including superpositions.
The Oxford team demonstrated ion pairing in this set-up for about 60 seconds, while the NIST/Washington team managed to keep theirs entangled for 1.5 seconds. That doesn’t sound like much, but that’s relatively stable when it comes to qubits.
Both teams confirm that their two atoms are entangled with a very high probability; 0.998 for one, 0.979 for the other (of a maximum of one),John Timmer reports for Ars Technica. “The NIST team even showed that it could track the beryllium atom as it changed state by observing the state of the magnesium atom.
Further, both teams were able to successfully perform a Bell test by using the logic gate to entangle the pairs of different-species ions, and then manipulating and measuring them independently.
[W]e show that quantum logic gates between different isotopic species are possible, can be driven by a relatively simple laser system, and can work with precision beyond the so-called ‘fault-tolerant threshold’ precision of approximately 99 percent – the precision necessary to implement the techniques of quantum error correction, without which a quantum computer of useful size cannot be built,said Lucas in an Oxford press release.
Of course, we don’t have proper quantum computers to actually test these components in the context of a functioning system – that will have to be the next step, and international teams of scientists and engineers are racing to get us there. We can’t wait to see it when they do.
The papers have been published in Nature here and here.
ORIGINAL: ScienceAlert
18 DEC 2015

Forward to the Future: Visions of 2045

By Hugo Angel,

DARPA asked the world and our own researchers what technologies they expect to see 30 years from now—and received insightful, sometimes funny predictions
Today—October 21, 2015—is famous in popular culture as the date 30 years in the future when Marty McFly and Doc Brown arrive in their time-traveling DeLorean in the movie “Back to the Future Part II.” The film got some things right about 2015, including in-home videoconferencing and devices that recognize people by their voices and fingerprints. But it also predicted trunk-sized fusion reactors, hoverboards and flying cars—game-changing technologies that, despite the advances we’ve seen in so many fields over the past three decades, still exist only in our imaginations.
A big part of DARPA’s mission is to envision the future and make the impossible possible. So ten days ago, as the “Back to the Future” day approached, we turned to social media and asked the world to predict: What technologies might actually surround us 30 years from now? We pointed people to presentations from DARPA’s Future Technologies Forum, held last month in St. Louis, for inspiration and a reality check before submitting their predictions.
Well, you rose to the challenge and the results are in. So in honor of Marty and Doc (little known fact: he is a DARPA alum) and all of the world’s innovators past and future, we present here some highlights from your responses, in roughly descending order by number of mentions for each class of futuristic capability:
  • Space: Interplanetary and interstellar travel, including faster-than-light travel; missions and permanent settlements on the Moon, Mars and the asteroid belt; space elevators
  • Transportation & Energy: Self-driving and electric vehicles; improved mass transit systems and intercontinental travel; flying cars and hoverboards; high-efficiency solar and other sustainable energy sources
  • Medicine & Health: Neurological devices for memory augmentation, storage and transfer, and perhaps to read people’s thoughts; life extension, including virtual immortality via uploading brains into computers; artificial cells and organs; “Star Trek”-style tricorder for home diagnostics and treatment; wearable technology, such as exoskeletons and augmented-reality glasses and contact lenses
  • Materials & Robotics: Ubiquitous nanotechnology, 3-D printing and robotics; invisibility and cloaking devices; energy shields; anti-gravity devices
  • Cyber & Big Data: Improved artificial intelligence; optical and quantum computing; faster, more secure Internet; better use of data analytics to improve use of resources
A few predictions inspired us to respond directly:
  • Pizza delivery via teleportation”—DARPA took a close look at this a few years ago and decided there is plenty of incentive for the private sector to handle this challenge.
  • Time travel technology will be close, but will be closely guarded by the military as a matter of national security”—We already did this tomorrow.
  • Systems for controlling the weather”—Meteorologists told us it would be a job killer and we didn’t want to rain on their parade.
  • Space colonies…and unlimited cellular data plans that won’t be slowed by your carrier when you go over a limit”—We appreciate the idea that these are equally difficult, but they are not. We think likable cell-phone data plans are beyond even DARPA and a total non-starter.
So seriously, as an adjunct to this crowd-sourced view of the future, we asked three DARPA researchers from various fields to share their visions of 2045, and why getting there will require a group effort with players not only from academia and industry but from forward-looking government laboratories and agencies:

Pam Melroy, an aerospace engineer, former astronaut and current deputy director of DARPA’s Tactical Technologies Office (TTO), foresees technologies that would enable machines to collaborate with humans as partners on tasks far more complex than those we can tackle today:
Justin Sanchez, a neuroscientist and program manager in DARPA’s Biological Technologies Office (BTO), imagines a world where neurotechnologies could enable users to interact with their environment and other people by thought alone:
Stefanie Tompkins, a geologist and director of DARPA’s Defense Sciences Office, envisions building substances from the atomic or molecular level up to create “impossible” materials with previously unattainable capabilities.
Check back with us in 2045—or sooner, if that time machine stuff works out—for an assessment of how things really turned out in 30 years.
# # #
Associated images posted on www.darpa.mil and video posted at www.youtube.com/darpatv may be reused according to the terms of the DARPA User Agreement, available here:http://www.darpa.mil/policy/usage-policy.
Tweet @darpa
[email protected]

Computer Learns to Write Its ABCs

By Hugo Angel,

Danqing Wang Computer ABC
Photo-illustration: Danqing Wang
A new computer model can now mimic the human ability to learn new concepts from a single example instead of the hundreds or thousands of examples it takes other machine learning techniques, researchers say.

The new model learned how to write invented symbols from the animated show Futurama as well as dozens of alphabets from across the world. It also showed it could invent symbols of its own in the style of a given language
.The researchers suggest their model could also learn other kinds of concepts, such as speech and gestures.

Although scientists have made great advances in .machine learning in recent years, people remain much better at learning new concepts than machines.

People can learn new concepts extremely quickly, from very little data, often from only one or a few examples. You show even a young child a horse, a school bus, a skateboard, and they can get it from one example,” says study co-author Joshua Tenenbaum at the Massachusetts Institute of Technology. In contrast, “standard algorithms in machine learning require tens, hundreds or even thousands of examples to perform similarly.

To shorten machine learning, researchers sought to develop a model that better mimicked human learning, which makes generalizations from very few examples of a concept. They focused on learning simple visual concepts — handwritten symbols from alphabets around the world.

Our work has two goals: to better understand how people learn — to reverse engineer learning in the human mind — and to build machines that learn in more humanlike ways,” Tenenbaum says.

Whereas standard pattern recognition algorithms represent symbols as collections of pixels or arrangements of features, the new model the researchers developed represented each symbol as a simple computer program. For instance, the letter “A” is represented by a program that generates examples of that letter stroke by stroke when the program is run. No programmer is needed during the learning process — the model generates these programs itself.

Moreover, each program is designed to generate variations of each symbol whenever the programs are run, helping it capture the way instances of such concepts might vary, such as the differences between how two people draw a letter.

The idea for this algorithm came from a surprising finding we had while collecting a data set of handwritten characters from around the world. We found that if you ask a handful of people to draw a novel character, there is remarkable consistency in the way people draw,” says study lead author Brenden Lake at New York University. “When people learn or use or interact with these novel concepts, they do not just see characters as static visual objects. Instead, people see richer structure — something like a causal model, or a sequence of pen strokes — that describe how to efficiently produce new examples of the concept.

The model also applies knowledge from previous concepts to speed learn new concepts. For instance, the model can use knowledge learned from the Latin alphabet to learn the Greek alphabet. They call their model the Bayesian program learning or BPL framework.

The researchers applied their model to more than 1,600 types of handwritten characters in 50 writing systems, including Sanskrit, Tibetan, Gujarati, Glagolitic, and even invented characters such as those from the animated series Futurama and the online game Dark Horizon. In a kind of .Turing test, scientists found that volunteers recruited via .Amazon’s Mechanical Turk had difficulty distinguishing machine-written characters from human-written ones.

The scientists also had their model focus on creative tasks. They asked their system to create whole new concepts — for instance, creating a new Tibetan letter based on what it knew about letters in the Tibetan alphabet. The researchers found human volunteers rated machine-written characters on par with ones developed by humans recruited for the same task.

We got human-level performance on this creative task,” study co-author Ruslan Salakhutdinov at the University of Toronto.

Potential applications for this model could include

  • handwriting recognition,
  • speech recognition,
  • gesture recognition and
  • object recognition.
Ultimately we’re trying to figure out how we can get systems that come closer to displaying human-like intelligence,” Salakhutdinov says. “We’re still very, very far from getting there, though.“The scientists detailed .their findings in the December 11 issue of the journal Science.


By Charles Q. Choi
Posted 10 Dec 2015 | 20:00 GMT