Google’s London-based AI outfit DeepMind has created two different types of AI that can use their ‘imagination’ to plan ahead and perform tasks with a higher success rate than AIs without imagination. Sorry if I made you click because you wanted AIs predicted flying cars. I promise this is cool too.
In a post on their site, DeepMind researchers give a short review of “a new family of approaches for imagination-based planning.” The so-called Imagination-Augmented Agents, or I2As, use an internal ‘imagination encoder’ that helps the AI decide what are and what aren’t useful predictions about its environment.
The researchers argue that giving AI imagination is crucial for dealing with real-world environments, where it’s helpful to test a few possible outcomes of actions ‘in your head’ to predict which one is best.
Recently, DeepMind’s founder Demis Hassabis wrote a paper published in Neuronabout how the development of general-purpose AI is dependent on understanding and encoding human abilities like imagination, curiosity, and memory into AI. With these papers, his company seems to be making headway in at least one of those areas.
The I2A ‘agents’ in the papers were tasked with different situations to test their predictive abilities, “including the puzzle game Sokoban and a spaceship navigation game.” Sokoban is a puzzle game in which a little alien has to push boxes into the right place – it can not pull though, so one wrong move can screw up the whole round.
To challenge the agent, the researchers had every level procedurally generated and only gave the agent one try to solve it, because “this encourages the agent to try different strategies ‘in its head’ before testing them in the real environment,” they wrote.
The agents ended up performing better than its imagination-less counterparts. They learned how to navigate the puzzles with less experience by extracting more information from their internal simulations. When the researchers added a ‘manager’ component that helped create a plan, it “learns to solve tasks even more efficiently with fewer steps.”
Of course the type of imagination described in these papers is nowhere near what humans are capable of, but it does show that AIs can and benefit from being able to efficiently imagine different scenarios before acting.
As Hassabis wrote in the Neuron paper, creating agents with an imagination that can rival what we can do “is perhaps the hardest challenge for AI research: to build an agent that can plan hierarchically, is truly creative, and can generate solutions to challenges that currently elude even the human mind.” But step by step, we might be getting there.
At the international GELP Moscow 2017 Summit, which was held in Moscow on 1st to 3rd of November, leaders of Russian and foreign education from 12 states outlined practical steps for systematical change of the global educational environment, taking into account the existing global challenges of civilization.
Global Venture Alliance took part in a closed round table session organized by Forbes Russia.
The authoritative international alliance Global Education Leaders` Partnership (GELP), which influences the development of school education abroad, for the first time, chose Russia as a meeting point for leaders and experts of the world educational community. The official operator of GELP is the Global Venture Alliance.
Global Venture Alliance is the general sponsor of the event on the occasion of the 100th anniversary of the magazine Forbes.
The Global Venture Alliance team on October 17 at the international forum "Open Innovations 2017" in Skolkovo organized a closed signing between RZhD, IBM, NES and NCC in the field of transport logistics.
Tim Cook has not-so-subtly hinted that Apple is working on some sort of augmented reality product. And while ARKit may be the start, a patent application published today hints at what Apple could be picturing down the road.
Apple has teamed up with Australian-based Cochlear to bring iPhone users the first made for iPhone Cochlear implant.
Approved by the U.S. Food and Drug Administration in June, Cochlear’s Nucleus 7 Sound Processor can now stream sound directly from a compatible iPhone, iPad or iPod touch to the sound processor.
Scientists at MIT are using Wi-Fi and AI to determine your emotional state. They’ve created an algorithm that can detect and measure individual heartbeats by bouncing RF signals off of people.
Yep. Glass is back.
Engineers will soon conduct a crucial test of a futuristic technology championed by entrepreneur Elon Musk that seeks to revolutionize transportation by sending passengers and cargo packed into pods through an intercity system of vacuum tubes.
Uber announced that it will be combining its rides-on-demand business and UberEATS, its food ordering and delivery business, in Russia and neighboring markets, with Yandex.Taxi, the ridesharing business.
Tesla will up its game in the renewable energy storage market with the largest lithium-ion battery storage facility in the world, to be built for Australia’s Hornsdale Wind Farm, with a completion date of December 1, 2017.
Amazon’s Alexa voice platform has now passed 15,000 skills — the voice-powered apps that run on devices like the Echo speaker, Echo Dot, newer Echo Show and others. The figure is up from the 10,000 skills Amazon officially announced back in February, which had then represented a 3x increase from September.
During official talks with his Brazilian counterpart, Michel Temer, in Moscow on Wednesday, Russian President Vladimir Putin hailed the cooperation between the two countries in the field of technologies.
Putin noted that Russia and Brazil already cooperate closely in space exploration and that there are plans to develop cooperation in Earth monitoring from space.
UCSB researchers have proposed a new method for 3D through-wall imaging that utilizes drones and WiFi