Say hello to the capsule network.
AI has enjoyed huge growth in the past few years, and much of that success is owed to deep neural networks, which provide the smarts behind impressive tricks like image recognition. But there is growing concern that some of the fundamental principles that have made those systems so successful may not be able to overcome the major problems facing AI—perhaps the biggest of which is a need for huge quantities of data from which to learn (for a deep dive on this, check out our feature "Is AI Riding a One-Trick Pony?").
Google’s Geoff Hinton appears to be among those fretting about AI's future. As Wired reports, Hinton has unveiled a new take on traditional neural networks that he calls capsule networks. In a pair of new papers—one published on the arXIv, the other on OpenReview—Hinton and a handful of colleagues explain how they work.
Their approach uses small groups of neurons, collectively known as capsules, which are organized into layers to identify things in video or images. When several capsules in one layer agree on having detected something, they activate a capsule at a higher level—and so on, until the network is able to make a judgment about what it sees. Each of those capsules is designed to detect a specific feature in an image in such a way that it can recognize them in different scenarios, like from varying angles.
Hinton claims that the approach, which has been in the making for decades, should enable his networks to recognize objects in new situations with less data than regular neural nets must use.
In the papers published so far, capsule networks have been shown to keep up with regular neural networks when it comes to identifying handwritten characters, and they make fewer errors when trying to recognize previously observed toys from different angles. But for the moment, at least, they're still a bit slower than their traditional counterparts.
Now, then, comes the interesting part. Will these systems provide a compelling alternative to traditional neural networks, or will they stall? We can expect the machine-learning community to implement the work, and fast, in order to find out. Either way, those concerned about the limitations of current AI systems can be heartened by the fact that researchers are pushing the boundaries to build new deep-learning alternatives.
Why Meta’s latest large language model survived only three days online
Galactica was supposed to help scientists. Instead, it mindlessly spat out biased and incorrect nonsense.
DeepMind’s game-playing AI has beaten a 50-year-old record in computer science
The new version of AlphaZero discovered a faster way to do matrix multiplication, a core problem in computing that affects thousands of everyday computer tasks.
A bot that watched 70,000 hours of Minecraft could unlock AI’s next big thing
Online videos are a vast and untapped source of training data—and OpenAI says it has a new way to use it.
The White House just unveiled a new AI Bill of Rights
It's the first big step to hold AI to account.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.