Inside Google’s secretive X laboratory, known for inventing self-driving cars and augmented-reality glasses, a small group of researchers began working several years ago on a simulation of the human brain.
There Google scientists created one of the largest neural networks for machine learning by connecting 16,000 computer processors, which they turned loose on the Internet to learn on its own.
Presented with 10 million digital images found in YouTube videos, what did Google’s brain do? What millions of humans do with YouTube: look for cats.
The neural network taught itself to recognize cats, which is actually no frivolous activity. This week, the researchers will present the results of their research at a conference in Edinburgh, Scotland.
The Google scientists and programmers will note that while it is hardly news that the Internet is full of cat videos, the simulation nevertheless surprised them. It performed far better than any previous effort by roughly doubling its accuracy in recognizing objects in a challenging list of 20,000 distinct items.
The research is representative of a new generation of computer science that is exploiting the falling cost of computing and the availability of huge clusters of computers in giant data centers. It is leading to significant advances in areas as diverse as machine vision and perception, speech recognition and language translation.
Although some of the computer science ideas that the researchers are using are not new, the sheer scale of the software simulations is leading to learning systems that were not previously possible. And Google researchers are not alone in exploiting the techniques, which are referred to as “deep learning” models. Last year, Microsoft scientists presented research showing that the techniques could be applied equally well to build computer systems to understand human speech.
“This is the hottest thing in the speech recognition field these days,” said Yann LeCun, a computer scientist who specializes in machine learning at the Courant Institute of Mathematical Sciences at New York University.
And then, of course, there are the cats.
To find them, the Google research team, led by the Stanford University computer scientist Andrew Ng and the Google fellow Jeff Dean, used an array of 16,000 processors to create a neural network with more than 1 billion connections. They then fed it random thumbnails of images, one each extracted from 10 million YouTube videos.
The videos were selected randomly, in itself an interesting comment on what interests humans in the Internet age. However, the research is also striking. That is because the software-based neural network created by the researchers appeared to closely mirror theories developed by biologists that suggest individual neurons are trained inside the brain to detect significant objects.
Currently, much commercial machine-vision technology is done by having humans “supervise” the learning process by labeling specific features. In the Google research, the machine was given no help in identifying features.
“The idea is that instead of having teams of researchers trying to find out how to find edges, you instead throw a tonne of data at the algorithm and you let the data speak and have the software automatically learn from the data,” Ng said.
“We never told it during the training: ‘This is a cat,’” said Dean, who originally helped Google design the software that lets it easily break programs into many tasks that can be computed simultaneously. “It basically invented the concept of a cat. We probably have other ones that are side views of cats.”
The Google brain assembled a dreamlike digital image of a cat by employing a hierarchy of memory locations to successively cull out general features after being exposed to millions of images. However, the scientists said that it appeared they had developed a cybernetic cousin to what takes place in the brain’s visual cortex.
Neuroscientists have discussed the possibility of what they call the “grandmother neuron,” specialized cells in the brain that fire when they are exposed repeatedly or “trained” to recognize a particular face of an individual.
“You learn to identify a friend through repetition,” said Gary Bradski, a neuroscientist at Industrial Perception, in Palo Alto, California.
While the scientists were struck by the parallel emergence of the cat images, as well as human faces and body parts in specific memory regions of their computer model, Ng said he was cautious about drawing parallels between his software system and biological life.
“A loose and frankly awful analogy is that our numerical parameters correspond to synapses,” Ng said.
He said that one difference was that despite the immense computing capacity that the scientists used, it was still dwarfed by the number of connections found in the brain.
“It is worth noting that our network is still tiny compared to the human visual cortex, which is 10-to-the-sixth-power times larger in terms of the number of neurons and synapses,” the researchers wrote.
Despite being dwarfed by the immense scale of biological brains, the Google research provides new evidence that existing machine-learning algorithms improve greatly as the machines are given access to large pools of data.
“The Stanford/Google paper pushes the envelope on the size and scale of neural networks by an order of magnitude over previous efforts,” Georgia Tech College of Computing executive director of high-performance computing David Bader said.
He said that rapid increases in computer technology would close the gap within a relatively short period of time: “The scale of modeling the full human visual cortex may be within reach before the end of the decade.”
Google scientists said that the research project had now moved out of the Google X laboratory and was being pursued in the division that houses the company’s search business and related services. Potential applications include improvements to image search, speech recognition and machine language translation.
Despite their success, the Google researchers remained cautious about whether they had hit upon the holy grail of machines that can teach themselves.
“It’d be fantastic if it turns out that all we need to do is take current algorithms and run them bigger, but my gut feeling is that we still don’t quite have the right algorithm yet,” Ng said.
Could Asia be on the verge of a new wave of nuclear proliferation? A look back at the early history of the North Atlantic Treaty Organization (NATO), which recently celebrated its 75th anniversary, illuminates some reasons for concern in the Indo-Pacific today. US Secretary of Defense Lloyd Austin recently described NATO as “the most powerful and successful alliance in history,” but the organization’s early years were not without challenges. At its inception, the signing of the North Atlantic Treaty marked a sea change in American strategic thinking. The United States had been intent on withdrawing from Europe in the years following
My wife and I spent the week in the interior of Taiwan where Shuyuan spent her childhood. In that town there is a street that functions as an open farmer’s market. Walk along that street, as Shuyuan did yesterday, and it is next to impossible to come home empty-handed. Some mangoes that looked vaguely like others we had seen around here ended up on our table. Shuyuan told how she had bought them from a little old farmer woman from the countryside who said the mangoes were from a very old tree she had on her property. The big surprise
The issue of China’s overcapacity has drawn greater global attention recently, with US Secretary of the Treasury Janet Yellen urging Beijing to address its excess production in key industries during her visit to China last week. Meanwhile in Brussels, European Commission President Ursula von der Leyen last week said that Europe must have a tough talk with China on its perceived overcapacity and unfair trade practices. The remarks by Yellen and Von der Leyen come as China’s economy is undergoing a painful transition. Beijing is trying to steer the world’s second-largest economy out of a COVID-19 slump, the property crisis and
Former president Ma Ying-jeou’s (馬英九) trip to China provides a pertinent reminder of why Taiwanese protested so vociferously against attempts to force through the cross-strait service trade agreement in 2014 and why, since Ma’s presidential election win in 2012, they have not voted in another Chinese Nationalist Party (KMT) candidate. While the nation narrowly avoided tragedy — the treaty would have put Taiwan on the path toward the demobilization of its democracy, which Courtney Donovan Smith wrote about in the Taipei Times in “With the Sunflower movement Taiwan dodged a bullet” — Ma’s political swansong in China, which included fawning dithyrambs