'Ex Machina', here we come: A new algorithm helps computers learn the way we do

In a 'visual Turing test,' human judges couldn't tell them apart

Machine learning is all about getting computers to "understand" new concepts, but it's still a pretty inefficient process, often requiring hundreds of examples for training. That may soon change, however, thanks to new research published on Friday.

Aiming to shorten the learning process and make it more like the way humans acquire and apply new knowledge based on just a few examples, a team of researchers has developed what they call a Bayesian Program Learning framework and then used it to teach computers to identify and reproduce handwritten characters based on just a single example.

Whereas standard pattern-recognition algorithms represent concepts as configurations of pixels or collections of features, the BPL approach learns by “explaining” the data provided to the algorithm -- in this case, the sample character. Concepts are represented as probabilistic computer programs and the algorithm essentially programs itself by constructing code to produce the letter it sees. It can also capture variations in the way different people draw a given letter.

The model also “learns to learn” by using knowledge from previous concepts to speed learning on new ones, so it can use knowledge of the Latin alphabet to learn letters in the Greek alphabet more quickly, for example.

Most compelling of all is that the algorithm allowed computers to pass a sort of "visual Turing test." Specifically, the researchers asked both humans and computers to reproduce a series of handwritten characters after being shown just a single example of each; in some cases, subjects were asked to create entirely new characters in the style of those originally shown. Bottom line: human judges couldn't tell the results apart.

The researchers have applied their model to more than 1,600 types of handwritten characters in 50 writing systems, including Sanskrit, Tibetan, Gujarati and Glagolitic. They even tried it on invented characters such as those from the television series "Futurama."

A paper describing the research was published Friday in the journal Science. Its authors were Brenden Lake, a Moore-Sloan Data Science Fellow at New York University; 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.

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

Join the newsletter!

Error: Please check your email address.
Show Comments

Blog Posts

Maintaining trust in a sceptical world: The power of brand trust

The faith people have in brands creates opportunity for those brands to become trusted advisors. In turn, this builds success by increasing the brand’s profile, letting it broaden its product offering and driving stronger customer loyalty.

Dan Ratner

managing director, uberbrand

When growth stalls: How to boost growth in large organisations

The push to start new businesses continues. In Q1 2017, the number of seed and angel deals increased by 1.4 per cent compared to Q1 2016.

Con Frantzeskos

CEO, Penso

Why we need diversity in marketing

​When we read articles about the need for increased diversity in marketing land, it is often through the lens of gender.

Jodie Sangster

CEO, ADMA

Interesting insight, well explained and the examples are just apt.Thanks for sharing!

FreshMindIdeas

The politics of branding - Brand science - CMO Australia

Read more

When the world that we live in floods with gigabytes of content every day, we have to learn to be selective about it. Such educational we...

Paulina Cameron

ADMA launches education program to tackle viewability, ad fraud and brand safety

Read more

Hi, i am an Aistralian ALK patient, been on xalkori dec 13 to oct 15 and achieved remission of disease, since been on Ceritinib until no...

gary packer

Pfizer Australia adopts AI-powered digital analyst tool for sales and marketing decision making

Read more

Hi James, shouldn't marketers also be focusing on collecting and utilizing up to date first-party profiling data on customers so that mes...

Tom

3 ways customer data can increase online sales conversion

Read more

Wouldn't reconnecting with younger consumers be in direct contravention of the code on alcohol advertising?

Tim Palmer

Vodka Cruiser reconnects with younger consumers via category-first Facebook Live campaign

Read more

Latest Podcast

Getting Intimate with CX Ep 5: Tammy Marshall, founder, The B Hive

How much of customer experience is having the foresight to know what those individuals might like, versus asking them? In Episode 5 of this new podcast series, BrandHook MD, Pip Stocks, talks with Tammy Marshall about the importance of asking your customers questions, how consistency plays a role in engagement, but how the unexpected adds extra value.

More podcasts

Sign in