New algorithm lets machines learn like humans

Scientists have created a new algorithm that captures human-like learning abilities. The algorithm enables computers to recognize and draw simple handwritten characters after exposure to just a few examples.

See Full Article

Their work, summarized in an article published Thursday in the journal Science, represents a major advancement in the field of machine learning.

The new algorithm, called “Bayesian Program Learning,” attempts to mimic the way humans learn new concepts. When humans are exposed to a new concept – say a new symbol or object– they can often recognize and understand the new concept after being exposed to it just a few times.

During a teleconference, U of T assistant computer science professor and article co-author Ruslan Salakhutdinov, gave the example of thumbs-up or high-five gestures that humans can recognize and perform as needed, after limited exposure to them.

While there are now machines and programs that can also learn to recognize symbols and patterns, for example object or speech recognition, these machines often depend on tens or hundreds of examples in order to function accurately.

But the new algorithm can recognize and learn a new concept after a limited number of examples, the authors of the study say. As well, the algorithm can perform more creative tasks beyond simple recognition. These tasks include generating new concepts, breaking down concepts into their respective parts, and understanding the relationship between those respective parts.

Salakhutdinov said the algorithm is a significant advancement in the field of artificial intelligence.

"It has been very difficult to build machines that require as little data as humans when learning a new concept," he said in a statement. "Replicating these abilities is an exciting area of research connecting machine learning, statistics, computer vision, and cognitive science."

Recognizing and reproducing handwritten characters

Using the algorithm, a computer was able to create a copy of handwritten characters, after being shown an example. In this case, the new characters were letters from different alphabets from around the world.

The computer was also able to mimic the way humans draw characters, by using similar pen strokes, stroke order and direction, the researchers said.

Finally, the computer was able to generate a totally new character when shown a set and asked to produce another one to be included in the set.

When the researchers compared the computer-generated results with characters produced by humans, they found them to be "mostly indistinguishable."

Machine learning

Can you tell the difference between humans and machines? Humans and machines were given an image of a novel character (top) and asked to produce new exemplars. The nine-character grids in each pair that were generated by a machine are (by row) B, A; A, B; A, B. (Photo courtesy Brenden Lake)

Machine learning

Can you tell the difference between humans and machines? Humans and machines were given an image of a novel character (top) and asked to produce new exemplars. The nine character grids in each pair that were generated by a machine are (by row) 1, 2; 2, 1; 1, 1. (Photo courtesy Brenden Lake)

Brenden Lake, the study's lead author and a Moore-Sloan Data Science Fellow at New York University, said the algorithm can lead to the development of better learning machines.

"Our results show that by reverse engineering how people think about a problem, we can develop better algorithms," he said in a statement. "Moreover, this work points to promising methods to narrow the gap for other machine learning tasks."

Joshua Tenenbaum, a professor at MIT in the Department of Brain and Cognitive Sciences and the Center for Brains, Minds and Machines and a co-author of the article, echoed the sentiment.

"I've wanted to build models of these remarkable abilities since my own doctoral work in the late nineties," he said in a statement. "We are still far from building machines as smart as a human child, but this is the first time we have had a machine able to learn and use a large class of real-world concepts – even simple visual concepts such as handwritten characters – in ways that are hard to tell apart from humans."

'Bayesian Program Learning'

The algorithm works by representing concepts as simple computer programs.

For example, the letter "A" is represented by computer code, which will generate examples of that letter when the code is run, NYU said in a statement released Thursday.

However, no human programmer is required during the learning process, rather, the algorithm programs itself by building code to produce the letter it sees. And, unlike standard computer programs that produce the same output every time, these special programs produce different outputs every time they’re run, the statement said.

This captures the variation that exists among concepts, such as the natural differences between how two people might draw the letter "A," the statement added.

In a video explaining his work, Lake said that while the algorithm only currently works for handwritten characters, it could eventually be used in other domains.

"The key point is that we need to learn the right form of representation, not just learning from bigger data sets, in order to build more powerful and more human-like learning patterns," he said.



Advertisements

Latest Tech & Science News

  • Extinction of world's biggest mammals tied to spread of humans

    Tech & Science CBC News
    The spread of humans around the world from Africa thousands of years ago wiped out big mammals in a trend that, if it continues, could make the cow the biggest mammal on Earth in a few centuries' time, a scientific study said on Thursday. Source
  • Trudeau not ready to join British PM's ban on single-use plastics

    Tech & Science CBC News
    Prime Minister Justin Trudeau stopped short today of echoing British Prime Minister Theresa May's call for Commonwealth members to ban single-use plastics — but pointed to a planned discussion at the next G7 summit, being hosted by Canada later this year. Source
  • Alberta university criticized over plan to bestow David Suzuki honorary degree

    Tech & Science CTV News
    EDMONTON - The University of Alberta is being criticized for its decision to give David Suzuki an honorary degree. The university announced earlier this month that the environmentalist will be one of 13 recipients in June. Source
  • Meet the newest 'exploding ant' that sacrifices itself for the good of the colony

    Tech & Science CBC News
    Deep in the forests of Borneo live species of ants with a rather novel way of fending off enemies: they explode. While they were first identified in 1916, no new species have been discovered since 1935. Now, a group of international scientists from multiple disciplines including botany, chemistry and entomology, have discovered 15 more separate species of these kamikaze ants. Source
  • Don't hold your breath: Secret behind super deep-diving ability revealed

    Tech & Science CBC News
    An age-old nomadic community of Southeast Asian boat-dwellers who get their food from the sea appear to have evolved enlarged spleens that may help explain their extreme diving prowess, a new study suggests. The spleen stores oxygen-rich red blood cells that it can release into the bloodstream, enabling divers to hold their breath for longer periods of time under water. Source
  • Put the toolbox away - new robot assembles IKEA chairs

    Tech & Science CTV News
    Sick of struggling with incomprehensible instructions and a baffling array of planks and screws? Help is at hand in the form of a new robot that can assemble an IKEA chair in minutes. The robot, developed by scientists at Singapore's Nanyang Technological University, can put together the wooden IKEA chair in just eight minutes and 55 seconds -- a swift timing that may give even DIY enthusiasts a run for their money. Source
  • Wreckage of warship sunk by Japanese torpedoes discovered off Solomon Islands

    Tech & Science CBC News
    A Microsoft co-founder's mission to locate sunken warships in the South Pacific has chalked up another victory with the discovery of the USS Helena nearly 75 years after it was sunk by Japanese torpedoes in the waters off the Solomon Islands, reviving stories of the battle-tested ship's endurance and the nearly unbelievable survival of 165 of the crewmen. Source
  • 'We are sorry': Facebook execs grilled by Canadian legislators amid Cambridge Analytica scandal

    Tech & Science CBC News
    Senior members of the Facebook leadership team faced a rough ride from MPs on Thursday for their failure to inform more than 600,000 Canadians that their privacy might have been compromised. For more than two years, Facebook knew that the personal information of potentially thousands of Canadians was in the hands of a third party — without their consent, and in contravention of Canadian privacy law. Source
  • Dashing to make a flight? Order food to your gate through a delivery app

    Tech & Science CTV News
    TORONTO -- P.J. Mastracchio is used to dashing through airports only to arrive at his gate and have a terrible feeling start to sink in: hunger pangs. The frequent business traveller's routine rush often leaves him with no time to grab a bite. Source
  • Windsor biologists using 3D printed robo-toads to study mating rituals in Costa Rica

    Tech & Science CBC News
    A research project at the University of Windsor is using 3D printed lookalikes to discover why a species of toad in Costa Rica turns bright yellow for a single day during its mating season. The project started a decade ago when Dan Mennill and Stephanie Doucet witnessed a remarkable transformation while studying birds in Costa Rica. Source