GOOGLE PLANS TO USE ARTIFICIAL INTELLIGENCE TO DO EVERYTHING
DeepMind is the artificial intelligence venture by Google and it has the potential to be able
to change the world.
Google has made an intelligence that can think on its own accord and they are hoping that
they will be able to integrate how humans think with machines.
DEEPMIND HAS HUGE COLLECTION OF BRAINS, RESOURCES AND UNLIMITED FUNDING
Google took over DeepMind, a London-based company, in 2014 and they have now turned
it into the biggest collection of both resources and brain power, with the focus being on developing
artificial intelligence.
The world is Google's oyster as they have more than 205 PhDs and unlimited resources
and funding.
The company has two goals, the first of which is to try to solve intelligence and the second
is they want to use that intelligence to do just about everything else.
The first step will see Google and DeepMind working out how much the brain in humans allowed
us to take over the planet.
They have more psychologists and neuroscientists, along with cognitive scientists, all hacking
at some of the best theories regarding how the human brain works and then disseminating
it to the computer scientists, who then try to replicate that in processors for artificial
intelligence.
IF SUCCESSFUL GOOGLE WOULD BE MOST POWERFUL ENTITY ON EARTH Step two is going to see them
applying this to every else, and yes in the literal form.
If they should be able to conquer this Google would essentially become the single most powerful
entity on planet Earth.
Google isn�t trying to hide this and they are not afraid to talk openly about the plans
to anyone that will listen to what they have to say.
The founder of DeepMind, Demis Hassabis, talks about the plan very clearly in a video.
In the video, he talks about what DeepMind has been doing over the last 12 to 18 months.
Demis admitted that they had the biggest collection of scientific brainpower anywhere in the world.
DeepMind should be thought of a hybrid he said, a mixture of some of the best academia
around the globe and the focus of some of the best start-ups from Silicon Valley.
Google helped to turbocharge the mission of the company and they want to understand intelligence
and then recreate it by artificial means and then go on to using this to solve just about
everything else, which Demis says is plausible.
DEEPMIND BUILDING GENERAL PURPOSE LEARNING SYSTEM Demis went on to say that they want
to build what would be the world�s first general-purpose learning system.
All systems they build learn to undertake tasks from raw data, they are not told anything
about the environment, and instead, they learn everything from first principals.
DeepMind had their first huge breakthrough a couple of years ago when they used virtual
environments, including games, to teach artificial intelligence.
They began with the old Atari games and built a system that was able to play any game from
pixel inputs, having to work out the rules of the game and how to win points.
The AI used trial and error along with the experience of the game.
ALPHA GO IS HUGE ACCOMPLISHMENT IN ARTIFICIAL INTELLIGENCE One of the most recent accomplishes
was with Alpha Go, a program that plays Go the board game, taking turns to place stones
on a board with the aim being to surround the opponent's stones, with the winner being
the one that captures the most stones.
While it sounds simple it is elegant and complex and taken extremely seriously in China, with
some playing it on a professional basis.
This game is very difficult for artificial intelligence to play, yet Alpha Go has managed
to master the game and won 4-1 when playing Lee Sedol, one of the top players in the world,
who has studied for many years.
If Google and DeepMind can design artificial intelligence that can learn the complex game
of GO and win, then who knows what they may be capable of achieving in the future.
See video clips for more info.
Links are in the article link below in
our description
Không có nhận xét nào:
Đăng nhận xét