Intelligence Too Big for a Single Machine

Since 1955, the AI field has gone through cycles of boundless optimism and sobering disillusion. Yet until recently, the supercomputer was the go-to operator of machine intelligence. But three forces have transformed that assumption in the last few years: the surge in data of all kinds, rapid progress in software to find patterns and insights in data, and advances in the technology of data processing, storage and communication.

Now, computing intelligence can be dispersed globally, marshaled and aggregated as necessary, from far-flung data centers in the digital cloud. Google led the way, showing the power of data-driven artificial intelligence delivered over the cloud, not only in search but also in tasks like language translation and computer vision. AI run through the cloud is now the dominant approach used by researchers at technology companies, universities and government labs.

More information:

http://bits.blogs.nytimes.com/2014/06/11/intelligence-too-big-for-a-single-machine/?_php=true&_type=blogs&_r=0

Mathematics of Robot Wrestling

Here’s how mathematicians might define the sport of wrestling. A system composed of two mechanical agents coupled via mechanical actions such as contact and collision. The aim of the contest is for one agent to floor the other while maintaining its own balance. The rest is just show business. That’s more or less exactly how researchers at Utsunomiya University in Japan describe the sport in developing a mathematical model of wrestling which they go on to test in a numerical simulation. The end result is a pair of autonomous mechanical wrestlers that compete to topple each other. Their model is simple in principle. Each wrestler is an inverted pendulum on a cart that can move backwards and forwards, a bit like balancing a pencil on your finger. These robot ‘wrestlers’ are joined at the tips by a spring that can stretch and compress. That means one wrestler can pull or push the other over. However, the opposing wrestler can take evasive action by moving in a way that stabilises itself and unbalances its opponent. The contest is over when one wrestler or the other falls to the ground. The question that researchers tackle is how best to design an intelligent controller that outperforms its opponent. The only action this controller can take is to move its cart backwards or forwards. Although simple in principle, this problem turns out to be hugely complex. In creating a mathematical model of the contest, researchers identify 17 different parameters that influence the behaviour of the wrestlers. These include the mass and length of the pendulum, the mass of the cart, acceleration due to gravity, the various properties of the spring, friction and so on. 

Each wrestler can end the bout in one of three configurations: standing up, having been pushed over or having been pulled over. So it’s not hard to see that there are nine possible outcomes in this contest. Of these, five permutations correspond to a draw, with both wrestlers having been pushed or pulled to the ground or with both remaining upright. The other four permutations correspond to a win for one side or the other. Each controller knows its position and the position of its opponent. It also knows how its own movement will produce a turning force that tends to unbalance the inverted pendulum. The question that the controller must solve is how to move in a way that maintains the upright position of its own pendulum while exerting a turning force that unbalances the opponent. To simplify matters, researchers assumed that the spring is rigid. But to avoid the trivial situation in which one controller simply drags the other over using brute force, they limit the impulse that each can produce. This turns the context into something of a chess game. One problem is that the solution space becomes so complex that the controllers cannot simulate it successfully and the contests end with the winner more or less chosen at random. But it turns out that when one controller has a short delay built in to its calculations; it becomes about twice as successful as an opponent that does not have this delay. That’s because undelayed controllers drive the system into complex states that they can no longer control but the delayed controllers never reach these levels of complexity and so turn out to be more successful. So far, all of this work has been pure numerical simulation but they have ambitious plans.

More information:

http://www.technologyreview.com/view/528086/the-complex-mathematics-of-robot-wrestling/

Beyond the Uncanny Valley

Graphics specialists are closing in on one of their field’s longest-standing and most sought-after prizes: an interactive and photo-realistically lifelike digital human. Such a digital double will change the way we think about actors, acting, entertainment, and computer games. In movies, digital doubles already replace human actors on occasion, sometimes just for moments, sometimes for most of a feature film. Within a decade or so, computer-game characters will be as indistinguishable from filmed humans as their movie counterparts. And in time, this capability will help bring movies and games together, and out of the union will come entirely new forms of entertainment. This blurring of the real and the digital became possible in movies recently when moviemakers reached a long-anticipated milestone: They crossed the ‘uncanny valley’. The term has been used for years to describe a problem faced by those using computer graphics to depict realistic human characters. When these creations stopped looking cartoonish and started approaching photo-realism, the characters somehow began to seem creepy rather than endearing. Some people speculated that the problem could never be solved; now it has proved to be just a matter of research and computing power.

Producing a fully realistic digital double is still fantastically expensive and time-consuming. It’s cheaper to hire even George Clooney than it is to use computers to generate his state-of-the-art digital double. However, the expense of creating a digital double is dependent on the costs of compute power and memory, and these costs will inevitably fall. Then, digital entertainment will enter a period of fast and turbulent change. An actor’s performance will be separable from his appearance; that is, an actor will be able to play any character—short, tall, old, or young. Some observers also foresee a new category of entertainment, somewhere between movies and games, in which a work has many plotted story lines and the viewer has some freedom to move around within the world of the story. Creators of entertainment also dream that technology will someday make itself invisible. Instead of painstakingly using cameras and computers to share their visions, they hope to be able to directly share the rich worlds of their imaginations using a form of electronic mind reading far more sophisticated than the brain-scan technology of today. That magic means not only being able to step inside a realistic virtual world without the cumbersome head-mounted screens that create virtual experiences today.

More information:

http://spectrum.ieee.org/computing/software/digital-actors-go-beyond-the-uncanny-valley

Robot People

As robotics quickly advance, scientists say the lines between robots and humans is beginning to blur. That means one day with robotic prosthetics that work seamlessly with a human's muscles, with tiny robots that swim in our blood streams and fix medical problems and nano-scale robots implanted in our brains, we will become robotic humans. As scary and sci-fi as that may sound, researchers say robotics will cure diseases, make amputees feel whole again and greatly extend our lives. Many, if not most people, will be wary of the idea of the melding of humans and robots, with images of Star Trek's evil cyborgs running through their heads. The fictional characters -- with both human and mechanical parts -- have superhuman strengths but have lost their individualism.

More than six years ago, a University of Arizona researcher who had successfully connected a moth's brain to a robot predicted that by 2017 or 2022 we'll be using hybrid computers that run a combination of technology and living organic tissue. Robotic exoskeletons have helped people suffering from paralysis walk again and the U.S. military is just weeks away from testing a new exoskeleton, or Iron Man-like suit, designed to make soldiers stronger, give them real-time battlefield information, monitor their vital signs and even stop their bleeding. Robotic prosthetics, using a built-in computer, 100 sensors and 17 motors can take natural cues from a user's residual limb, giving him or her the dexterity and grace to play a piano.

More information:

http://www.computerworld.com/s/article/9248923/Hail_cyborgs_The_line_between_robots_and_humans_is_blurring?pageNumber=1

Microsoft - 3D Audio

Just as a new generation of virtual reality goggles for video games are about to hit the market, researchers at Microsoft have come up with what could be the perfect accompaniment—a way for ordinary headphones to create a realistic illusion of sound coming from specific locations in space. 

In combination with a virtual reality device like the Oculus Rift, the new system could be used to make objects or characters in a virtual world sound as well as look like they are at a specific point in space, even if that is outside a person’s field of view. Researchers refer to the technology as 3D audio.

More information:

http://www.technologyreview.com/news/527826/microsofts-3-d-audio-gives-virtual-objects-a-voice/

DARPA's Augmented Reality

Six years ago, the Defense Advanced Research Projects Agency (DARPA) decided that they had a new dream. The agency wanted a system that would overlay digital tactical information right over the top of the physical world. So, they created a program called Urban Leader Tactical Response, Awareness and Visualization (ULTRA-Vis) to develop a novel and sophisticated augmented reality system for use by soldiers. Through half a decade and with the help of several military contractors, they succeeded. 

The program developed and integrated a light-weight, low-power holographic see-through display with a vision-enabled position and orientation tracking system. Using the ULTRA-Vis system, a Soldier can visualize the location of other forces, vehicles, hazards and aircraft in the local environment even when these are not visible to the Soldier. The system can be also used to communicate to the Soldier a variety of tactically significant information including imagery, navigation routes, and alerts.

More information:

http://www.theatlantic.com/technology/archive/2014/05/hows-darpas-augmented-reality-software-works/371652/

Computers of the Future

Computing experts at Sandia National Laboratories have launched an effort to help discover what computers of the future might look like, from next-generation supercomputers to systems that learn on their own — new machines that do more while using less energy. For decades, the computer industry operated under Moore’s Law, named for Intel Corp. co-founder Gordon Moore, who in 1965 postulated it was economically feasible to improve the density, speed and power of integrated circuits exponentially over time. But speed has plateaued, the energy required to run systems is rising sharply and industry can’t indefinitely continue to cram more transistors onto chips.

The plateauing of Moore’s Law is driving up energy costs for modern scientific computers to the point that, if current trends hold, more powerful future supercomputers would become impractical due to enormous energy consumption. Solving that conundrum will require new computer architecture that reduces energy costs, which are principally associated with moving data, Leland said. Eventually, computing also will need new technology that uses less energy at the transistor device-level. Sandia is well positioned to work on future computing technology due to its broad and long history in supercomputers, from architecture to algorithms to applications.

More information:

https://share.sandia.gov/news/resources/news_releases/beyond_moore/#.U4iaRy_GL9g

Printable Robots

Printable robots are those that can be assembled from parts produced by 3D printers and have long been a topic of research at MIT. Researchers demonstrate the promise of printable robotic components that, when heated, automatically fold into prescribed 3D configurations.

The initial work is a system that takes a digital specification of a 3D shape and generates the 2D patterns that would enable a piece of plastic to reproduce it through self-folding. In their later work, researchers present designs for resistors, inductors, and capacitors, as well as sensors and actuators.

More information:

http://www.sciencedaily.com/releases/2014/05/140530124420.htm