Skip to main content

Why teaching robots to play hide-and-seek could be the key to next-gen A.I.

 

Artificial general intelligence, the idea of an intelligent A.I. agent that’s able to understand and learn any intellectual task that humans can do, has long been a component of science fiction. As A.I. gets smarter and smarter — especially with breakthroughs in machine learning tools that are able to rewrite their code to learn from new experiences — it’s increasingly widely a part of real artificial intelligence conversations as well.

Recommended Videos

But how do we measure AGI when it does arrive? Over the years, researchers have laid out a number of possibilities. The most famous remains the Turing Test, in which a human judge interacts, sight unseen, with both humans and a machine, and must try and guess which is which. Two others, Ben Goertzel’s Robot College Student Test and Nils J. Nilsson’s Employment Test, seek to practically test an A.I.’s abilities by seeing whether it could earn a college degree or carry out workplace jobs. Another, which I should personally love to discount, posits that intelligence may be measured by the successful ability to assemble Ikea-style flatpack furniture without problems.

One of the most interesting AGI measures was put forward by Apple co-founder Steve Wozniak. Woz, as he is known to friends and admirers, suggests the Coffee Test. A general intelligence, he said, would mean a robot that is able to go into any house in the world, locate the kitchen, brew up a fresh cup of coffee, and then pour it into a mug.

As with every A.I. intelligence test, you can argue about how broad or narrow the parameters are. However, the idea that intelligence should be linked to an ability to navigate through the real world is intriguing. It’s also one that a new research project seeks to test out.

Building worlds

“In the last few years, the A.I. community has made huge strides in training A.I. agents to do complex tasks,” Luca Weihs, a research scientist at the Allen Institute for AI, an artificial intelligence lab founded by the late Microsoft co-founder Paul Allen, told Digital Trends.

AI2-Thor Tasks
Allen Institute for A.I.

Weihs cited DeepMind’s development of A.I. agents that are able to learn to play classic Atari games and beat human players at Go. However, Weihs noted that these tasks are “frequently detached” from our world. Show a picture of the real world to an A.I. trained to play Atari games, and it will have no idea what it is looking at. It’s here that the Allen Institute researchers believe they have something to offer.

The Allen Institute for A.I. has built up something of a real estate empire. But this isn’t physical real estate, so much as it is virtual real estate. It’s developed hundreds of virtual rooms and apartments — including kitchens, bedrooms, bathrooms, and living rooms — in which A.I. agents can interact with thousands of objects. These spaces boast realistic physics, support for multiple agents, and even states like hot and cold. By letting A.I. agents play in these environments, the idea is that they can build up a more realistic perception of the world.

Allen Institute for A.I.

“In [our new] work, we wanted to understand how A.I. agents could learn about a realistic environment by playing an interactive game within it,” Weihs said. “To answer this question, we trained two agents to play Cache, a variant of hide-and-seek, using adversarial reinforcement learning within the high-fidelity AI2-THOR environment. Through this gameplay, we found that our agents learned to represent individual images, approaching the performance of methods requiring millions of hand-labeled images — and even began to develop some cognitive primitives often studied by [developmental] psychologists.”

Rules of the game

Unlike regular hide-and-seek, in Cache, the bots take turns hiding objects such as toilet plungers, loaves of bread, tomatoes, and more, each of which boast their own individual geometries. The two agents — one a hider, the other a seeker — then compete to see if one can successfully hide the object from the other. This involves a number of challenges, including exploration and mapping, understanding perspective, hiding, object manipulation, and seeking. Everything is accurately simulated, even down to the requirement that the hider should be able to manipulate the object in its hand and not drop it.

Using deep reinforcement learning — a machine learning paradigm based on learning to take actions in an environment to maximize reward — the bots get better and better at hiding the objects, as well as seeking them out.

“What makes this so difficult for A.I.s is that they don’t see the world the way we do,” Weihs said. “Billions of years of evolution has made it so that, even as infants, our brains efficiently translate photons into concepts. On the other hand, an A.I. starts from scratch and sees its world as a huge grid of numbers which it then must learn to decode into meaning. Moreover, unlike in chess, where the world is neatly contained in 64 squares, every image seen by the agent only captures a small slice of the environment, and so it must integrate its observations through time to form a coherent understanding of the world.”

A.I. Hide and Seek Dynamic Experiment Results
Allen Institute for A.I.

To be clear, this latest work isn’t about building a supe-intelligent A.I. In movies like Terminator 2: Judgment Day, the Skynet supercomputer achieves self-awareness at precisely 2.14 a.m. Eastern Time on August 29, 1997. Notwithstanding the date, now almost a quarter century in our collective rearview mirror, it seems unlikely that there will be such a precise tipping point when regular A.I. becomes AGI. Instead, more and more computational fruits — low-hanging and high-hanging — will be plucked until we finally have something approaching a generalized intelligence across multiple domains.

Hard stuff is easy, easy stuff is hard

Researchers have traditionally gravitated toward complex problems for A.I. to solve based on the idea that, if the tough problems can be sorted, the easy ones shouldn’t be too far behind. If you can simulate the decision-making of an adult, can ideas like object permanence (the idea that objects still exist when we can’t see them) that a child learns within the first few months of its life really prove that difficult? The answer is yes — and this paradox that, when it comes to A.I., the hard stuff is frequently easy, and the easy stuff is hard, is what work such as this sets out to address.

“The most common paradigm for training A.I. agents [involves] huge, manually labeled datasets narrowly focused to a single task — for instance, recognizing objects,” said Weihs. “While this approach has had great success, I think it is optimistic to believe that we can manually create enough datasets to produce an A.I. agent that can act intelligently in the real world, communicate with humans, and solve all sorts of problems that it hasn’t encountered before. To do this, I believe we will need to let agents learn the fundamental cognitive primitives we take for granted by letting them freely interact with their world. Our work shows that using gameplay to motivate A.I. agents to interact with and explore their world results in them beginning to learn these primitives — and thereby shows that gameplay is a promising direction away from manually abeled datasets and towards experiential learning.”

A paper describing this work will be presented at the upcoming 2021 International Conference on Learning Representations.

Luke Dormehl
Former Digital Trends Contributor
I'm a UK-based tech writer covering Cool Tech at Digital Trends. I've also written for Fast Company, Wired, the Guardian…
The future of A.I.: 4 big things to watch for in the next few years
brain with computer text scrolling artificial intelligence

A.I. isn’t going to put humanity on the scrap heap any time soon. Nor are we one Google DeepMind publication away from superintelligence. But make no mistake about it: Artificial intelligence is making enormous strides.

As noted in the Artificial Intelligence Index Report 2021, last year the number of journal publications in the field grew by 34.5%. That’s a much higher percentage than the 19.6% seen one year earlier. A.I. is going to transform everything from medicine to transportation, and there are few who would argue otherwise.

Read more
Algorithmic architecture: Should we let A.I. design buildings for us?
Generated Venice cities

Designs iterate over time. Architecture designed and built in 1921 won’t look the same as a building from 1971 or from 2021. Trends change, materials evolve, and issues like sustainability gain importance, among other factors. But what if this evolution wasn’t just about the types of buildings architects design, but was, in fact, key to how they design? That’s the promise of evolutionary algorithms as a design tool.

While designers have long since used tools like Computer Aided Design (CAD) to help conceptualize projects, proponents of generative design want to go several steps further. They want to use algorithms that mimic evolutionary processes inside a computer to help design buildings from the ground up. And, at least when it comes to houses, the results are pretty darn interesting.
Generative design
Celestino Soddu has been working with evolutionary algorithms for longer than most people working today have been using computers. A contemporary Italian architect and designer now in his mid-70s, Soddu became interested in the technology’s potential impact on design back in the days of the Apple II. What interested him was the potential for endlessly riffing on a theme. Or as Soddu, who is also professor of generative design at the Polytechnic University of Milan in Italy, told Digital Trends, he liked the idea of “opening the door to endless variation.”

Read more
Emotion-sensing A.I. is here, and it could be in your next job interview
man speaking into phone

I vividly remember witnessing speech recognition technology in action for the first time. It was in the mid-1990s on a Macintosh computer in my grade school classroom. The science fiction writer Arthur C. Clarke once wrote that “any sufficiently advanced technology is indistinguishable from magic” -- and this was magical all right, seeing spoken words appearing on the screen without anyone having to physically hammer them out on a keyboard.

Jump forward another couple of decades, and now a large (and rapidly growing) number of our devices feature A.I. assistants like Apple’s Siri or Amazon’s Alexa. These tools, built using the latest artificial intelligence technology, aren’t simply able to transcribe words -- they are able to make sense of their contents to carry out actions.

Read more