Quoc Le lived in a small town in Vietnam without electricity.
This fact about Le's life tends to astonish him, especially since he is the leader of the «Google Brain» team, responsible for designing the artificial intelligence that today supports the work of millions of people in the world.
At 32, one of Le's main responsibilities at Google is to develop a neuralgic network in computers like that of human beings, or at least simulate it.
Is about build an artificial brain. But it is possible?
The origin
Le's first idea may not have been a machine with a human resemblance and a thought of its own. However, the concept of "intelligence" was always on Le's mind from childhood.
He humbly recalls the small technological advances seen in his hometown of Thuy Dong: the first television, the first car he saw by, the first rice cooker his family bought.
"At that time, each of those events in which technology entered our environment produced a sudden change in our lives," recalls Le.
Nevertheless, it was a photo of the moon landing in 1969 that inspired him to become a pioneer in artificial intelligence.
“I wondered: we are not the fastest animals on Earth, we cannot even fly, but we made it to the Moon. What is the only characteristic that we have that animals do not have?
Then I realized that it was all due to our brains. His intelligence".
At that time he assumed that intelligent machines were trivial, then he would understand that they were not.
"I decided I would build one," he remembers with a smile on his lips.
The investigation
He began researching artificial intelligence while studying at university in Australia and later during his PhD at Stanford University in the USA.
He was frustrated because the software of that time required a significant degree of human participation.
“The machines we described as intelligent required us to spend a lot of time coding decisions for the software to work. I wanted everything to be automatic", it states.
So he started working on a program that would allow the machine to think for itself.
Deep learning
The next step was to enter the field of deep learning, an attempt to make machines learn based on a group of algorithms.
It was already a developing field when it came to him, but he wanted the advancement to be faster thanks to the large-scale construction of a nerve system capable of processing a greater volume of data.
One of these artificial neurons - as Le describes it - reacted "very happy" to seeing a photo of a cat without even telling it what a cat was or without being given images labeled "cat" in advance.
Definitely it was an extraordinary advance, by showing that machines could learn at a very high level of precision, without the continuous human participation supplying information.
Another study showed a conversation between a machine and a human being, in which the machine was able to respond after studying the subtitles of the movies.
Human-machine dialogue
A study led by Le on the machine learning process collects the following dialogue between a computer and a human, to measure general knowledge:
– Human: Who is Skywalker?
- Machine: A hero.
– Human: Who is Bill Clinton?
- Machine: A millionaire.
– Human: Is the sky blue or black?
- Machine: Blue.
– Human: Do cats have tails?
- Machine: Yes.
– Human: Do cats have wings?
- Machine: No.
Google's brain
These findings and other results achieved in various studies have been incorporated into Google services, providing predictable responses to emails, recognizing places and people in photos, and performing searches with voice recognition.
Although deep learning is anticipated to be one of the most promising areas of study in the field of artificial intelligence, there are many other equally interesting areas.
Even within Google there is a team created after the acquisition of the British company DeepMind, in 2014, that uses machine learning technology for video games.
Last week the computer managed to defeat the champion of the ancient Chinese game known as Go, which has taken time to master due to the huge range of options it allows.
Other companies such as Facebook, Microsoft and Baidu (from China) have announced that they will invest in deep learning and other forms of artificial intelligence.
The future of Le
In Le's vision, at some point his technology will be transformed into artificial intelligence like the one shown in the movie "She" (Her, directed by Spike Jonze in 2013), in which operating systems act as personal assistants to human beings.
«What matters to me is build a machine that is capable of seeing, hearing and understanding each other"Says Le, although he admits that something like this will happen in a very distant future.
What this young Vietnamese has achieved in the last five years is to help make artificial intelligence available to everyone, and throughout the world, including the inhabitants of his hometown in Vietnam.
No Comment