Scientists teach machines to learn like humans

December 10, 2015
This paper compares human and machine learning for a wide range of simple visual concepts, or handwritten characters selected from alphabets around the world. This is an artist's interpretation of that theme. Art by Danqing Wang. Credit: Danqing Wang

A team of scientists has developed an algorithm that captures our learning abilities, enabling computers to recognize and draw simple visual concepts that are mostly indistinguishable from those created by humans. The work, which appears in the latest issue of the journal Science, marks a significant advance in the field—one that dramatically shortens the time it takes computers to 'learn' new concepts and broadens their application to more creative tasks.

"Our results show that by reverse engineering how people think about a problem, we can develop better algorithms," explains Brenden Lake, a Moore-Sloan Data Science Fellow at New York University and the paper's lead author. "Moreover, this work points to promising methods to narrow the gap for other machine learning tasks."

The paper's other authors were Ruslan Salakhutdinov, an assistant professor of Computer Science at the University of Toronto, and Joshua Tenenbaum, a professor at MIT in the Department of Brain and Cognitive Sciences and the Center for Brains, Minds and Machines.

When humans are exposed to a new concept—such as new piece of kitchen equipment, a new dance move, or a new letter in an unfamiliar alphabet—they often need only a few examples to understand its make-up and recognize new instances. While machines can now replicate some pattern-recognition tasks previously done only by humans—ATMs reading the numbers written on a check, for instance—machines typically need to be given hundreds or thousands of examples to perform with similar accuracy.

The video will load shortly.
This movie shows 20 different people drawing a novel character (left) and the algorithm predicting how those images were drawn (right). Credit: Brenden Lake

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

Salakhutdinov helped to launch recent interest in learning with 'deep neural networks,' in a paper published in Science almost 10 years ago with his doctoral advisor Geoffrey Hinton. Their algorithm learned the structure of 10 handwritten character concepts—the digits 0-9—from 6,000 examples each, or a total of 60,000 training examples.

In the work appearing in Science this week, the researchers sought to shorten the learning process and make it more akin to the way humans acquire and apply new knowledge—i.e., learning from a small number of examples and performing a range of tasks, such as generating new examples of a concept or generating whole new concepts.

The video will load shortly.
Credit: Brenden Lake

To do so, they developed a 'Bayesian Program Learning' (BPL) framework, where concepts are represented as simple computer programs. For instance, the letter 'A' is represented by computer code—resembling the work of a computer programmer—that generates examples of that letter when the code is run. Yet no programmer is required during the : the algorithm programs itself by constructing code to produce the letter it sees. Also, unlike standard computer programs that produce the same output every time they run, these probabilistic programs produce different outputs at each execution. This allows them to capture the way instances of a concept vary, such as the differences between how two people draw the letter 'A.'

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. Credit: Brenden Lake

While standard pattern recognition algorithms represent concepts as configurations of pixels or collections of features, the BPL approach learns "generative models" of processes in the world, making learning a matter of 'model building' or 'explaining' the data provided to the algorithm. In the case of writing and recognizing letters, BPL is designed to capture both the causal and compositional properties of real-world processes, allowing the algorithm to use data more efficiently. The model also "learns to learn" by using knowledge from previous concepts to speed learning on new concepts—e.g., using knowledge of the Latin alphabet to learn letters in the Greek alphabet. The authors applied their model to over 1,600 types of handwritten characters in 50 of the world's writing systems, including Sanskrit, Tibetan, Gujarati, Glagolitic—and even invented characters such as those from the television series Futurama.

In addition to testing the algorithm's ability to recognize new instances of a concept, the authors asked both humans and computers to reproduce a series of handwritten characters after being shown a single example of each character, or in some cases, to create new characters in the style of those it had been shown. The scientists then compared the outputs from both humans and machines through 'visual Turing tests.' Here, human judges were given paired examples of both the human and machine output, along with the original prompt, and asked to identify which of the symbols were produced by the computer.

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. Credit: Brenden Lake

While judges' correct responses varied across characters, for each visual Turing test, fewer than 25 percent of judges performed significantly better than chance in assessing whether a machine or a human produced a given set of symbols.

"Before they get to kindergarten, children learn to recognize new concepts from just a single example, and can even imagine new examples they haven't seen," notes Tenenbaum. "I've wanted to build models of these remarkable abilities since my own doctoral work in the late nineties. 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."

Explore further: Microsoft open sources Distributed Machine Learning Toolkit for more efficient big data research

More information: "Human-level concept learning through probabilistic program induction," by B.M. Lake et al. Science, www.sciencemag.org/lookup/doi/ … 1126/science.aab3050

Related Stories

Are computers creative?

December 8, 2015

High-tech automation is eliminating jobs like Pac-Man gobbling yellow dots. According to a 2013 study by the Oxford Martin School, 47 percent of U.S. jobs are susceptible to takeover by machines in the coming decades. Whether ...

Recommended for you

New method analyzes corn kernel characteristics

November 17, 2017

An ear of corn averages about 800 kernels. A traditional field method to estimate the number of kernels on the ear is to manually count the number of rows and multiply by the number of kernels in one length of the ear. With ...

Optically tunable microwave antennas for 5G applications

November 16, 2017

Multiband tunable antennas are a critical part of many communication and radar systems. New research by engineers at the University of Bristol has shown significant advances in antennas by using optically induced plasmas ...

3 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

FredJose
2 / 5 (4) Dec 11, 2015
Here you have highly intelligent agents use sophisticated tools and methods to perform rudimentary actions that humans do so easily. Yet some would have us believe that the human's abilities developed from pure chance and natural selection over many eons of time.
Except they have zero proof that that was the case. They can only whip it out of their heads and fling it at the public at large with the stern admonishment that anyone who doesn't believe it is an idiot that denies science. Just who is the un-scientific nut-case in this scenario? Just who is it that glibly discards the rigorous scientific method when it comes to the question of how things originated in the past?
antialias_physorg
5 / 5 (2) Dec 11, 2015
Except they have zero proof that that was the case.

You mean besides all the libraries, cave paintings and whatnot? Or are those all imaginary?

Here you have highly intelligent agents use sophisticated tools and methods to perform rudimentary actions that humans do so easily. Yet some would have us believe that the human's abilities developed from pure chance and natural selection over many eons of time.

Did you even read the article? It says right there:
For instance, the letter 'A' is represented by computer code—resembling the work of a computer programmer—that generates examples of that letter when the code is run. Yet NO PROGRAMMER is required during the learning process: THE ALGORITHM PROGRAMS ITSELF by constructing code to produce the letter it sees.
(caps mine)

This pretty much refutes what you're trying to claim.

You are the idiot here for failing to understand the science. It'd be less embarassing if you wouldn't display it so blatantly.
snerdguy
not rated yet Dec 14, 2015
It is advancements in A.I. like this that has me convince that our current, NASA driven ambitions of sending human beings into deep space to land on asteroids, to build a Mars colony or even a moon colony is a dangerous and costly dead end. The real explorers of such a hostile environment will be intelligent machines that can be built for the environment they are working in and that will be capable of reasoning and making decisions. They won't need human beings to drive them. The fragile humans can stay on Earth and ask them questions. They will learn from their environment just like we do and report back. Because they operate autonomously, they will explore vast expanses and gather large amounts of data in a much shorter time. This is our future in space, not some overpriced capsule.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.