Google’s DeepMind says it’s close to achieving ‘human-level’ artificial intelligence

DeepMind, a British company owned by Google, may be on the verge of achieving human-level artificial intelligence (AI).

Nando de Freitas, Research Scientist at DeepMind and Professor of Machine Learning at Oxford University, said it was ‘game over’ in terms of solving the toughest challenges in the race to achieve Artificial General Intelligence (AGI).

Artificial general intelligence refers to a machine or software that has the ability to understand or learn any intellectual task that a human can do without training.

According to De Freitas, the search for scientists is now seeking to scale AI programs, such as augmenting data and computing power, to create AGI.

Earlier this week, DeepMind unveiled a new AI “agent” called Gato that can complete 604 different tasks “across a wide range of environments.”

Gato uses a single neural network – a computing system with interconnected nodes that work like neurons in the human brain.

It can chat, comment on pictures, stack blocks with a real robot arm and even play the 80s Atari home video game console, DeepMind claims.

Scroll down for the video

DeepMind, a British company owned by Google, may be on the verge of achieving human-level artificial intelligence (file photo)

DeepMind, a British company owned by Google, may be on the verge of achieving human-level artificial intelligence (file photo)

Gato uses a single neural network — computing systems with interconnected nodes that act like neurons in the human brain — to complete 604 tasks, according to DeepMind.

Gato uses a single neural network — computing systems with interconnected nodes that act like neurons in the human brain — to complete 604 tasks, according to DeepMind.

artificial general intelligence

Artificial General Intelligence (AGI) is the ability of an intelligent agent to understand or learn any intellectual task that a human can perform.

Some commentators believe that we are decades away from achieving artificial general intelligence, and some doubt that we will see artificial general intelligence in this century.

Artificial General Intelligence has already been identified as a future threat that could wipe out humanity either on purpose or by accident.

De Freitas’ comments came in response to an opinion piece published in Next Web That humans alive today will never achieve AGI at all.

De Freitas tweeted: “It’s all about scale now! Game over! It’s about making these models bigger, safer, more computing efficient, faster…”

He acknowledged, however, that humanity is still far from creating an artificial intelligence that can pass the Turing Test – a test of a machine’s ability to demonstrate intelligent behavior that is equivalent to or indistinguishable from human behavior.

Following DeepMind’s announcement of Gato, The Next Web article said it doesn’t explain AGI more than virtual assistants like Amazon’s Alexa and Apple’s Siri, which are already on the market and in people’s homes.

“The Gato’s multitasking ability is much more like a video game console that can store 600 different games, rather than a game you can play 600 different ways,” said Tristan Green, a contributor to Next Web.

It’s not artificial general intelligence, it’s a neatly assembled, pre-trained set of tight models.

Gato is built to fulfill a variety of hundreds of missions, but this ability could jeopardize the quality of each mission, according to other commentators.

De Freitas tweeted:

De Freitas tweeted: “It’s all about scale now! Game over! It’s about making these models bigger, safer, more computing efficient, faster…”

In another opinion article, ZDNet Columnist Tiernan Ray wrote that the agent “actually isn’t great at many tasks.”

“On the other hand, the software is able to do a better job than custom machine learning software at controlling Sawyer’s robotic arm that stacks blocks,” Ray said.

On the other hand, it produces photo captions that are in many cases very poor quality.

“Her ability in standard chat dialogue with a human interlocutor is similarly mediocre, at times eliciting contradictory and illogical statements.”

For example, when chatting on a bot, Gato erroneously said that Marseille is the capital of France.

Also, a caption created by Gato to accompany a photo says “Man holding a banana to take a picture of it,” even though the man wasn’t holding the bread.

DeepMind Gato details in a new research paper titled “General Agent,” published on Arxiv prepress server.

The company’s authors said such an agent would show a “significant performance improvement” when scaled up.

Artificial General Intelligence has already been identified as a future threat that could wipe out humanity either on purpose or by accident.

Imagine conversations with Gattu when you were asked to be a chatbot.  One critic described Gato's ability to have a conversation with a human

Imagine conversations with Gattu when you were asked to be a chatbot. One critic described Gato’s ability to hold a conversation with a “mid-level” human being.

Earlier this week, British company DeepMind revealed Gato, a program that can chat, comment on photos, stack blocks with a real robot arm and even power the 80s home video game console Atari.  Some of the tasks that Gato was tested on were demonstrated in the DeepMind promo

Earlier this week, British company DeepMind revealed Gato, a program that can chat, comment on photos, stack blocks with a real robot arm and even power the 80s home video game console Atari. Some of the tasks that Gato was tested on were demonstrated in the DeepMind promo

Dr Stuart Armstrong of the University of Oxford’s Institute for the Future of Humanity said previously that AI will eventually render humans redundant and wipe us out.

He believes that machines will operate at speeds unimaginable to the human mind and will go beyond communication with humans to control the economy, financial markets, transportation, healthcare, and more.

Dr. Armstrong said that simple instructions for artificial general intelligence to “prevent human suffering” could be interpreted by a supercomputer as “killing all humans,” because human language is easily misinterpreted.

Before his death, Professor Stephen Hawking told BBC: “The development of full artificial intelligence could mean the end of the human race.”

During his lifetime, the famous British astrophysicist Professor Stephen Hawking (pictured) said that artificial intelligence

During his lifetime, the famous British astrophysicist Professor Stephen Hawking (pictured) said that artificial intelligence “could mean the end of the human race”

In a 2016 paper, DeepMind researchers acknowledged the need for a “big red button” to prevent a machine from completing a “malicious sequence of actions”.

DeepMind, founded in London in 2010 before being acquired by Google in 2014, is best known for creating artificial intelligence software that beat professional Go player Lee Sedol, the world champion, in a five-match match in 2016.

In 2020, the company announced that it had solved a 50-year-old problem in biology, known as the ‘protein folding problem’ — knowing how a protein’s amino acid sequence determines its three-dimensional structure.

DeepMind claimed to have solved the problem with 92 percent accuracy by training a neural network containing 170,000 known protein sequences and their different structures.

The company is perhaps best known for its AlphaGo AI software that beat professional Go player Lee Sedol, the world champion, in a five-match match.  Pictured, Go World Champion Lee Sedol of South Korea is seen before the first Google DeepMind Challenge Match against Google AlphaGo in March 2016

The company is perhaps best known for its AlphaGo AI software that beat professional Go player Lee Sedol, the world champion, in a five-match match. Pictured, Go World Champion Lee Sedol of South Korea is seen before the first Google DeepMind Challenge Match against Google AlphaGo in March 2016

What is Google’s Artificial Intelligence In-Depth Project?

DeepMind was founded in London in 2010 and acquired by Google in 2014.

It now has additional research centers in Edmonton and Montreal, Canada, and a DeepMind Applied team in Mountain View, California.

DeepMind is on a mission to push the boundaries of artificial intelligence, developing software that can learn to solve any complex problem without having to learn how to do it.

If successful, the company believes this will be one of the most important and useful scientific developments ever.

The company made headlines for a number of its creations, including software it created that taught itself how to play and win 49 completely different Atari titles, using only raw pixels as input.

For the first time in the world, AlphaGo has captured and won the world’s best player in G, one of the most complex and intuitive games of all time, with more locations than the atoms of the universe – and it won.

.

Leave a Comment