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

  • EPA keeps scientists from speaking about report on climate

    Tech & Science CTV News
    PROVIDENCE, R.I. -- The U.S. Environmental Protection Agency kept three scientists from speaking at a Monday event in a move condemned by researchers and Democratic members of Congress as an attempt by the agency to silence a discussion of climate change. Source
  • 12 big cities sign 'fossil-fuel-free streets' declaration

    Tech & Science CBC News
    Zero emissions areas could mean more parks, pedestrian areas or roads where only electric or hydrogen-powered vehicles could enter to make cities more attractive places to live. They did not define how big "major areas" would be. Source
  • Dandelions found in oilsands tailings could help clean them up: researchers

    Tech & Science CTV News
    A dandelion growing in the middle of a barren patch of oilsands tailings might unlock one way to help clean them up. Researchers at the University of Saskatchewan in Saskatoon have found the dandelion was hosting a fungus that ate the residual chemicals in the tailings. Source
  • 238 cities, regions want to house Amazon's 2nd headquarters

    Tech & Science Toronto Sun
    NEW YORK — Amazon said today that it received 238 proposals from cities and regions in the United States, Canada and Mexico hoping to be the home of the company’s second headquarters. The online retailer kicked off its hunt for a second home base in September, promising to bring 50,000 new jobs and spend more than US$5 billion on construction. Source
  • Stephen Hawking's PhD thesis goes online, website crashes

    Tech & Science CTV News
    LONDON -- Cambridge University has put Stephen Hawking's doctoral thesis online, triggering such interest that it crashed the university's website. Completed in 1966 when Hawking was 24, "Properties of Expanding Universes" explores ideas about the origins of the universe that have resonated through the scientist's career. Source
  • Britain to give Canada the shipwrecks of explorer Franklin

    Tech & Science CTV News
    LONDON - Britain will give Canada the shipwrecks of British explorer John Franklin, who tried to chart the Northwest Passage through the Canadian Arctic in 1845. The Ministry Of Defence said in a statement it would transfer the ownership of the HMS Erebus and HMS Terror to Parks Canada, but retain a small sample of artifacts. Source
  • N.S. vineyards flourish, for now, as temperatures grow warmer

    Tech & Science CTV News
    HALIFAX -- The skies have been brilliant blue in Nova Scotia wine country this October, the vines heavy with grapes, and winemakers like Sean Sears are seeing crops they could only vaguely hope for in the past. Source
  • 'No threat': Civil liberties groups support Chelsea Manning's bid to cross Canadian border

    Tech & Science CBC News
    A Canadian lawyer representing Chelsea Manning has formally requested the Canadian government reconsider its decision to block the American leaker of classified documents from entering the country. In support, more than 40 civil liberties groups, human rights advocates, privacy scholars, and academics have sent letters to the Canadian government highlighting Manning's advocacy for government transparency, prisoners' rights, and the LGTBQ community. Source
  • British writer launches seventh expedition to search for elusive pink-headed duck

    Tech & Science CTV News
    WASHINGTON -- Hope is the thing with feathers, poet Emily Dickinson wrote. For Richard Thorns, the feathers are pink. Thorns' hope? To prove that a colorful duck is not extinct. This week, he launches a seventh expedition into the inaccessible wilds of Myanmar to search for the pink-headed duck that hasn't been seen alive since 1949, and that was in India. Source
  • New chemical cocktails found in Hudson Bay polar bears

    Tech & Science CTV News
    New tests have found a wide range of previously undiscovered contaminants in polar bears around Hudson Bay. The new study complicates our understanding of the complex cocktail of chemicals the bears are exposed to as they try to adapt to changing climate, said Environment Canada researcher Robert Letcher. Source