In 2018, Google DeepMind Imbued A.I. With Human-Like Sight and Imagination

Humans can do a lot with a little context. If you see a picture of a toilet, you’ll know it’s probably flanked by a bathtub and sink. The picture didn’t include any of that stuff, but the human brain has a knack for filling in missing pieces. And now thanks to Google DeepMind computer scientists, so does artificial intelligence.

In a paper published in Science Magazine in June, the company described how it created a Generative Query Network (GQN) that can see and imagine almost like a human. In an accompanying video, research scientist S. M. Ali Eslami explains how the algorithm used “something akin to imagination” to turn a few two-dimensional images of a virtual room into a fleshed-out, three-dimensional environment.

This is #1 on Inverse’s list of the 20 Ways A.I. Became More Human in 2018.

“Our ability to learn about the world by simply looking at it is simply incredible. One of the biggest open problems in A.I. is figuring out what is necessary to allow computers to do the same,” he told Inverse in a written statement. “In this work, we train a neural network to predict what a scene might look like from new viewpoints.”

Article continues below

GIF of GQN agent operating in partially observed maze environments.

This A.I. was brought to life using a two-part system. The “representation network” translates the sample images into code the computer can understand. Next, the “generation network” creates everything else that isn’t shown in the initial images.

GIF of GQN agent “imagining” new viewpoints in rooms with multiple objects..

Like some of the other A.I. breakthroughs in 2018, this Deep Mind A.I.’s ability to extrapolate is a key step toward voice assistants that not only serve our needs, up anticipate them. Instead of having to command them, they’ll use context to know when to hand us a coffee or cook us a meal.

This was the landmark A.I. breakthrough of 2018 because it effectively made artificial intelligence significantly more intelligent across the board.A.I.

A new year means a potentially all-new look for AirPods. Apple’s buds are due for a refresh, having launched in late 2016, and a steady drumbeat of reliable rumors, analyst predictions, and unearthed patents have already painted a picture of what the AirPods 2 could look like.

On Tuesday, the United States Patent and Trademark Office granted Apple another series of 24 patents, one of which sketched out what appears to be a rather comprehensive overhaul to the AirPods charging case. This new design also comes with two new wireless charging features. The documents depict a case that can cordlessly charge itself and another device simultaneously, like, say, an Apple Watch.

SpaceX’s Falcon 9 has aided the completion of one of the largest tech upgrades in history. The company’s Iridium-8 launch, completed on Sunday, was both the first mission of 2019 and the last mission of a two-year project to upgrade a satellite constellation. After the event, CEO Elon Musk shared images of the rocket after successfully landing back on Earth.

The CES “booth babes” may largely be a thing of the past, and in 2019, cute-as-heck robots have seemingly replaced them, dominating the show floor.

Among the cute robots of CES are Sony’s very good (and very expensive) robo-pup, Aibo, and the doe-eyed, sloth-like Lovot Groove X, which you can see in a video below. But as many of CES 2019’s most adorable bots illustrate, robo-cuteness is more than just an aesthetics trend — it’s a prediction.

Duplex, Google’s next-generation voice assistant took voice assistants to the next level. As of late November, Pixel phones can be commanded to call a restaurant and reserve a table using the Google Assistant.

The software update was announced in March during the annual Google I/O developer conference. It gave the Google Assistant a shockingly realistic human voice, which Inverse got to experience first-hand in June. During the live demo, the A.I. managed mimic the “uhhs” and “umms” that pepper human conversation, an incredible leap in the field of natural language processing.