Artificial Intelligence With True Intelligence
Deep Learning is a method of Artificial Intelligence that could be generalizable to many kinds of applications. Artificial Intelligence is finally getting smart.
At the beginning of the movie A.I. (Artificial Intelligence), 2001, a character states that, “To create an artificial being has been the dream of man since the birth of science.” Well, there is probably an element of truth to this. We, or at least some filmmakers, always have the fascination to create intelligent machines to (pardon me for saying) serve our laziness.
But if in the movie human found themselves in the brink of war with intelligent robots called the Mache, in reality many computer scientists agreed that even in the next 40 years it still very difficult for computers to have the raw processing power of human brain.
So when Ray Kurzweil approached Google founder and CEO, Larry Page, what he has in mind is ideas about how to build a truly intelligent machine, one that could understand language and then make decisions on its own. Just like “smart” human will do. It later became obvious to Ray that for such effort he would need Google-scale data and computing power.
“I could try to give you some access to it,” Page told Kurzweil. “But it’s going to be very difficult to do that for an independent company.” So instead providing him with the data he requested, Page suggested that Kurzweil join Google. Kurzweil officially become Google’s Director of Engineering early this year. “Joining Google in January is literally the culmination of my 50 years focus on Artificial Intelligence,” said Kurzweil.
What made Kurzweil who never worked in another companies agreed to join Google is not only the company’s computing resources, but also the startling progress the company has made in a branch of Artificial Intelligence, the branch called deep- learning system. The intelligent macchine’s software attempts to mimic the activity in layers of neurons in the neocortex. The wrinkly 80 percent of the brain where thinking occurs. The deep-learning software learns, in a very real sense, to recognize patterns in digital presentations of sounds, images and other data.
The basic idea of deep-learning system is that software can simulate the neocortex’s large array of neurons in an artificial “neural network”. Artificial Intelligence scientists have focused on the development in decades but only led to many dissappointments. Thanks to the improvements in mathematical formulas and increasing developments of intelligent machines, computer scientists can now model many more layers of virtual neurons than ever before.
With a greater depth than ever, computer scientists are producing remarkable advances in speech and image recognition. In June last year, a Google deep-learning system that had been shown 10 million images from YouTube videos proved almost twice as good as any previous image recognition effort at identifying objects such as cats.
Google has also used the technology to cut the error rate on speech recognition in its latest Android mobile software. In October 2012, Microsoft chief research officer Rick Rashid wowed attendees at a lecture in China with a demonstration of speech software that transcribed his spoken words into English text with an error rate of 7 percent, translated them into Chinese-language text, and then simulated his own voice uttering them in Mandarin.
Google, the company that started by the Stanford’s alumni, Sergey Brin and Larry Page, has become magnet for intelligent machine’s development like Artificial Intelligence. In March, the company reportedly bought a startup cofounded by Geoffrey Hinton, a university of Toronto computer science professor. Hinton plans to take ideas out of the ‘field’ and apply them to real problems such as image recognition, search and natural-language understanding.
Both Artificial Intelligence researchers, Kurzweil and Hinton, are hopeful that intelligent machines may finally escape the pages of science fiction. Intelligent machines is indeed, starting to transfrom everything from communications and computing to medicine, manufacturing and transportation. The possibilities are apparent in IBM’s Watson computer which uses some deep-learning techniques and is now being trained to help doctors make better decisions. Microsoft has deployed deep- learning system in its Windows Phone and Bing voice search.
However, to extend deep-learning sytem into applications beyond speech and image recognition will require more conceptual and software breakthrough, not to mention many more advances in processing power. We probably won’t see intelligent machines that can think for themselves for years, or perhaps decades, but for now deep-learning system has overcome some of the grand challenges in Artificial Intelligence.
With deep-learning system, the prospects of Artificial Intelligence that truly intelligent are intriguing. Better image search clearly would help YouTube, for instance. “Deep learning” model is claimed capable of using phoneme data from English to more quickly train systems to recognize the spoken sounds in other languages. Whilst sophisticated image recognition could make Google’s self-driving cars much better.
Kurzweil has long had a vision of intelligent machines. While in high school, he wrote software that enabled a computer to create original music in various classical styles, which he later demonstrated in a 1965 appearance on the TV show I’ve Got a Secret. Since then, Kurzweil inventions have included several firsts such as: a print-to-speech reading machine, software that could scan and digitize printed text in any font, music synthesizers that could re-create the sound of orchestral instruments, and a speech recognition system with a large vocabulary. Kurzweil believes that computers would assist humans far more effective if they could reliably recognize patterns and make inferences about the world.
He envisions a “cybernetic friend” that listens in on your phone conversations, reads your e-mail, and tracks your every move—if you let it, of course—so it can tell you things you want to know even before you ask. But this is not his immediate goal at Google. For now, Kurzweil aims to help computers understand and even speak in natural language. “My mandate is to give computers enough understanding of natural language to do useful thing like to do a better job of search or to do a better job of answering questions,” he says. Although Kurzweil’s vision of intelligent machines is still years from reality, deep-learning system is likely to spur other applications beyond speech and image recognition in the near future.
(Date: 14 June 2013; Frida)