Of course, we're not yet living in the era of fully aware machines: This thing isn't perfect yet. Though the machine identified 20,000 objects, doubling accuracy, it's still nothing compared to the complexity of the human brain. "It is worth noting that our network is still tiny compared to the human visual cortex, which is a million times larger in terms of the number of neurons and synapses,” the researchers wrote. But, once these computers get good enough, can't we imagine some sort of privacy-violating computer overlord? This is the stated end-goal, after-all. "The focus of this work is to build high-level, classspecic feature detectors from unlabeled images. For instance, we would like to understand if it is possible to build a face detector from only unlabeled images," explains the research.
Don't let the cats distract from the eerie and impressive parts of Google's latest research. Out of the super-secret Google X Labs, researchers have gotten an artificial brain to identify objects—in this trollish case, cats—without giving it any information. "The idea is that instead of having teams of researchers trying to find out how to find edges, you instead throw a ton of data at the algorithm and you let the data speak and have the software automatically learn from the data," researcher Andrew Y. Ng told The New York Times' John Markoff. This computer didn't know anything about these images at all, yet—and here's the scary part—it could identify these cats without any information. "Experimental results using classification and visualization confirm that it is indeed possible to build high level features from unlabeled data. In particular, using a hold-out test set consisting of faces and distractors, we discover a feature that is highly selective for faces," the researchers write. If people didn't like facial recognition software before, this development should make them even more wary.
This article is from the archive of our partner The Wire.