Artificial intelligence (abbreviated AI, also some times called Synthetic Intelligence) is defined as intelligence exhibited by an artificial entity. Such a system is generally assumed to be a computer.

AI forms a vital branch of computer science, dealing with intelligent behavior, learning and adaptation in machines. Research in AI is concerned with producing machines to automate tasks requiring intelligent behavior. Examples include control, planning and scheduling, the ability to answer diagnostic and consumer questions, handwriting, speech, and facial recognition. As such, it has become an engineering discipline, focused on providing solutions to real life problems. AI systems are now in routine use in economics, medicine, engineering and the military, as well as being built into many common home computer software applications, traditional strategy games like computer chess and other video games.

For topics relating specifically to full human-like intelligence, see Strong AI and science fiction.

More on [ Artificial intelligence ]

Robotics / artificial intelligence

Flight of fancy
Thu, 03 Dec 2009 05:00:00 -0000
In its first 18 years, the Association for Unmanned Vehicle Systems International’s annual aerial-robotics competition posed four successive challenges, which robotics researchers had to meet using entirely autonomous aerial vehicles — no remote control allowed. The first challenge, which stood for three years, was to move a metal disc from one end of an arena to another. The fourth challenge was to travel three kilometers and find a way into a specific building: it stood for eight years. But this summer, for the first time in the competition’s history, a challenge fell in its first year, to a team of students representing MIT’s Robust Robotics Group.The competition presented a scenario mimicking the aftermath of a nuclear meltdown. The aerial robot had to navigate its way through a window and into a maze simulating the hallways of an evacuated building, locate the control room, identify a gauge ostensibly indicating radiation levels, photograph it, and transmit the photo to a base station over a radio connection. Unlike the fourth challenge, the fifth denied the vehicles access to GPS data.Kyle Snyder, AUVSI’s senior technical director, said the MIT team’s feat came as “a pleasant surprise” to members of his organization and to other industry experts.“I talked to some of the industry folks that attended the competition,” Snyder says, “and they said that there’s no way they could have done what the MIT and Georgia Tech teams were doing. Especially the MIT team — to actually pull together the sensors, the platform, and the understanding of what it was going to take to complete that mission. There’s nobody out there that could have done it.”Competitors could use any type of aerial vehicle they chose, although, of course, it had to be small enough to operate indoors. Georgia Tech used a helicopter with two parallel rotors, and Embry-Riddle Aeronautical University used an innovative vehicle with a single spinning wing, like the pod of a maple seed. But the MIT team, which consisted of graduate students Abraham Bachrach, Ruijie He, and Sam Prentice and undergrads Anton de Winter and Garrett Hemann, used a battery-powered, off-the-shelf robot called a quad helicopter. The quad helicopter — or “quad” for short — is about two feet across and has a rotor at each of its four corners.According to Nicholas Roy, the associate professor in the Department of Aeronautics and Astronautics who directed the students’ work, arming the quad with the software necessary to navigate a hallway required addressing “a fundamental research question.” The quad’s information about its immediate environment comes from a laser rangefinder that shoots out beams of light and gauges how long their reflections take to return. Since the lasers that emit the light lie in the same plane, they allow the quad to construct a two-dimensional map of its surroundings — a cross-section as seen from above. But the quad is continually buffeted by disturbances in the air, and if it tilts slightly, the map can change dramatically.The quad, however, is also equipped with gyros and accelerometers that can measure its motion in three dimensions, so its twisting and tilting can be correlated with the changes in its environmental map. The MIT team developed algorithms that can use those correlations to give the quad some three-dimensional information about its surroundings.Because the twisting and tilting points the rotors in unanticipated directions, it causes the quad to drift, so the algorithms also had to be fast: they had to be able to build their maps before the quad’s position changed too drastically. But in robotic control systems, gains in processing speed usually come at the expense of accuracy. To figure out exactly how much accuracy they could afford to give up, the MIT researchers tested their system in the Computer Science and Artificial Intelligence Laboratory’s motion capture studio, a large room with regularly spaced cameras along the tops of its walls. They were thus able to compare the quad’s own sense of its position with very precise external measurements, and they determined that, if the quad’s onboard computer was performing well, it could gauge its position with an accuracy of about five centimeters — exactly the margin of error that it had when passing through the window at the beginning of the competition.The location-mapping software ran onboard the quad’s own processor, but to assemble a higher-level map of the entire maze, the quad radioed its measurements to a nearby base station that ran what Roy calls the “planning algorithm.” “One of our successes has been a planning algorithm that takes into account the fact that the sensor is limited,” Roy says. He points out, for instance, that the laser rangefinder has a 120-degree blind spot and a range of only 30 meters. “When we fly down long hallways,” Roy says, “the hallway may be longer than the maximum range. So down the corridor you see nothing, and you can build up a lot of speed very quickly without realizing it.” The planning algorithm thus keeps the quad oriented so that the rangefinder’s blind spot is directed at one of the side walls, so the quad can gauge its velocity by reference to the back wall — or any other obstacle with a fixed position — until the approaching wall comes into view.Eric Johnson, the faculty advisor to the Georgia Tech team that entered the competition, says that while the MIT team developed a “fantastic system,” “there’s plenty of work to be done to make that kind of system practical and usable.” He points out, for instance, that “there’s a lot that can be done to make the system more robust and faster,” and that “another big detail to tackle is the 3-D aspect of it: although their system certainly could handle some aspects in three dimensions, I don’t think it had what would be necessary to, say, go up and down stairs.”Roy agrees that “to do more three-dimensional operations — to be able to find a desk and land on a desk — the camera is clearly the right kind of sensor for that.” In fact, he says, the quad that completed the AUVSI challenge was equipped to process data from its camera, “but we ended up not using it because we did not need it,” he says. “In order to minimize points of failure, you turn off the things you don’t need.” But in his group’s ongoing research, Roy says, “we are moving more and more toward integrating the camera and the laser.”
From nature, robots
Fri, 25 Sep 2009 04:00:00 -0000
To a robot designer like Sangbae Kim, the animal kingdom is full of inspiration."I always look at animals and ask why they are the way they are," says Kim, an assistant professor of mechanical engineering at MIT. "As an engineer, looking at them and speculating is fascinating."While a graduate student at Stanford, Kim drew inspiration from the gecko to build a climbing robot, and he is now designing a running robot that mimics the movements of a cheetah. Such agile, fast-moving robots could perform military surveillance and search-and-rescue missions deemed too dangerous for humans to undertake.His Biomimetic Robotics Lab is one of several at MIT pursuing biologically inspired engineering. A team of mechanical engineers has built robotic fish, and materials scientists have designed moisture-collecting materials that mimic a beetle's shell. Evolution has produced finely tuned adaptations over millions of years, so it only makes sense to turn to nature for design ideas. However, while Kim seeks inspiration in nature, he's not trying to produce exact robotic copies of a particular animal. Such copying would be difficult to achieve and not necessarily the most effective design strategy."There are millions of things that animals have to adapt for, and it is almost impossible to compare evolution to our engineering/mathematical optimization process," says Kim. "And you have to be careful about copying other features that may not be related to the particular function you want to achieve. Therefore, extracting scientific principle is extremely important for designers like me."StickybotWhen Kim and his colleagues at Stanford set out to build a climbing robot, at first they figured they needed to make the robot's feet sticky. However, they soon realized that very sticky feet can't detach very easily.Their approach shifted dramatically with the 2006 discovery, by Lewis and Clark College biologist Kellar Autumn, that geckos use a phenomenon called directional adhesion to stick to walls."The gecko gave us a completely new perspective. Stickiness does not necessarily come from chemical composition; it can come from mechanical properties and geometry," says Kim. "The geometry enables strange phenomena such as directional adhesion, which sticks in only one direction."The pads of a gecko's feet are covered with a forest of tiny hairs called setae, some of which are one-twentieth the width of a human hair. The setae, in turn, branch into hundreds of tiny smaller hairs called spatulae, which are about one-thousandth the width of a human hair. These hairs cling to surfaces using tiny molecular interactions known as van der Waals forces. Collectively, the forces are strong enough to support the gecko's weight as it scrambles up a vertical surface.To demonstrate, Kim rummages around in a desk drawer in his office and pulls out a small rectangle of the gecko-inspired adhesive material, which resembles a tiny patch of blue Astroturf. A compact disc gently held against the horizontal surface attaches securely in one direction and then easily detaches in the opposite direction. The adhesive is covered with hairs made of rubber silicone, which are thicker than those on a gecko's paw (about four times thicker than a human hair). Because thicker hairs require smoother surfaces for adhesion, Stickybot can only climb extremely smooth surfaces like glass.Kim and his colleagues, led by Stanford professor Mark Cutkosky, first demonstrated Stickybot in 2006, and Time magazine named it one of that year's best inventions. The paper describing the robot also won the 2008 Best Paper Award for the IEEE Transactions on Robotics.Potential applications for the stickybot technology include exterior repair of underwater oil pipelines and window washing. Kim also plans to start designing climbing equipment for humans using the directional adhesion technology.Need for speedKim, who arrived at MIT as an assistant professor in June, is now turning his attention to a speedier robot, inspired by the cheetah. Four graduate students have just begun working on the cheetah project, and within the next two years Kim hopes to have a prototype that can run 35 miles per hour.Though his design incorporates principles from a variety of running animals, including horses and dogs, Kim zeroed in on the cheetah because of its special adaptations for speed. One feature he plans to mimic is the flexibility of the cheetah's backbone, which gives extra speed or force to its running motion.To demonstrate how extra joints can add force and speed, Kim leans back in his chair and mimics throwing a baseball, in slow motion — first the shoulder, then the elbow, then the wrist bend. The force imparted by each of those joints adds up, allowing a pitcher to throw a faster pitch. In the same way, the joints of the cheetah's leg — hip, knee and ankle — are aided by the extra speed generated by its bending backbone, which is much more flexible than that of other running mammals.Kim and his students plan to start building and testing prototypes within the next 18 months, after using a computer model to calculate the optimal limb length and weight, gait and torque of the hip and knee joints. He expects that the biggest challenge will be getting enough power out of the motor to furnish the desired speed. To that end, he plans to build the robot out of lightweight carbon fiber-foam composite, so less power is needed to propel it. Another difficult problem is coordinating the control of three joints in four legs. Those 12 joints each have to move in concert with the others, and they need to be able to react smoothly to disturbances in the gait, such as tripping over a rock, and regain balance.Kim believes his robots could be a significant improvement over current wheeled robots used for scouting and search and rescue, which are efficient but slow. "It's going to be very exciting to see how fast we can go and how rough a terrain we can navigate."
Fish and chips
Mon, 24 Aug 2009 05:00:00 -0000
Borrowing from Mother Nature, a team of MIT researchers has built a school of swimming robo-fish that slip through the water just as gracefully as the real thing, if not quite as fast.Mechanical engineers Kamal Youcef-Toumi and Pablo Valdivia y Alvarado have designed the sleek robotic fish to more easily maneuver into areas where traditional underwater autonomous vehicles can't go. Fleets of the new robots could be used to inspect submerged structures such as boats and oil and gas pipes; patrol ports, lakes and rivers; and help detect environmental pollutants."Given the (robotic) fish's robustness, it would be ideal as a long-term sensing and exploration unit. Several of these could be deployed, and even if only a small percentage make it back there wouldn't be a terrible capital loss due to their low cost," says Valdivia y Alvarado, a recent MIT PhD recipient in mechanical engineering.Robotic fish are not new: In 1994, MIT ocean engineers demonstrated Robotuna, a four-foot-long robotic fish. But while Robotuna had 2,843 parts controlled by six motors, the new robotic fish, each less than a foot long, are powered by a single motor and are made of fewer than 10 individual components, including a flexible, compliant body that houses all components and protects them from the environment. The motor, placed in the fish's midsection, initiates a wave that travels along the fish's flexible body, propelling it forward.The robofish bodies are continuous (i.e., not divided into different segments), flexible and made from soft polymers. This makes them more maneuverable and better able to mimic the swimming motion of real fish, which propel themselves by contracting muscles on either side of their bodies, generating a wave that travels from head to tail."Most swimming techniques can be copied by exploiting natural vibrations of soft structures," says Valdivia y Alvarado.As part of his doctoral thesis, Valdivia y Alvarado created a model to calculate the optimal material properties distributions along the robot's body to create a fish with the desired speed and swimming motion. The model, which the researchers initially proposed in the journal Dynamic Systems Measurements and Control (ASME), also takes into account the robot's mass and volume. A more detailed model is described in Valdivia y Alvarado's thesis and will soon be published along with new applications by the group. Other researchers, including a team at the University of Essex, have developed new generations of robotic fish using traditional assembly of rigid components to replicate the motions of fish, but the MIT team is the only one using controlled vibrations of flexible bodies to mimic biological locomotion."With these polymers, you can specify stiffness in different sections, rather than building a robot with discrete sections," says Youcef-Toumi. "This philosophy can be used for more than just fish" — for example, in robotic prosthetic limbs.Mimicking fishWith motors in their bellies and power cords trailing as they swim, the robo-fish might not be mistaken for the real thing, but they do a pretty good fish impersonation. The team's first prototypes, about five inches long, mimic the carangiform swimming technique used by bass and trout. Most of the movement takes place in the tail end of the body. Fish that use this type of motion are generally fast swimmers, with moderate maneuverability.Later versions of the robo-fish, about eight inches long, swim like tuna, which are adapted for even higher swimming speeds and long distances. In tuna, motion is concentrated in the tail and the peduncle region (where the tail attaches to the body), and the amplitude of body motions in this region is greater than in carangiform fish.Real fish are exquisitely adapted to moving through their watery environment, and can swim as fast as 10 times their body length per second. So far, the MIT researchers have gotten their prototypes close to one body length per second - much slower than their natural counterparts but faster than earlier generations of robotic fish.The new robo-fish are also more durable than older models — with their seamless bodies, there is no chance of water leaking into the robots and damaging them. Several four-year-old prototypes are still functioning after countless runs through the testing tank, which is filled with tap water.Current prototypes require 2.5 to 5 watts of power, depending on the robot's size. That electricity now comes from an external source, but in the future the researchers hope to power the robots with a small internal battery.Later this fall, the researchers plan to expand their research to more complex locomotion and test some new prototype robotic salamanders and manta rays."The fish were a proof of concept application, but we are hoping to apply this idea to other forms of locomotion, so the methodology will be useful for mobile robotics research — land, air and underwater — as well," said Valdivia y Alvarado.The work was funded by the Singapore-MIT Alliance and Schlumberger Ltd.
Quanta Computer extends collaboration with CSAIL
Fri, 17 Jul 2009 05:00:00 -0000
Quanta Computer is extending its research collaboration with MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) for another five years. As part of the collaboration, the company is devoting an additional $25 million to work with CSAIL on developing new information technologies."Through Project T-Party, Quanta Computer has become an extraordinary partner for MIT, inspiring and supporting our research at the very frontiers of mobile computing," said MIT President Susan Hockfield. "Given our shared passion for inventing the future, we enter this new phase of our relationship with high hopes for equally important results."The original five-year collaboration, the $20 million T-Party Project, was launched in 2005 as a joint research effort between Quanta Computer, Inc., the largest notebook computer company in the world, and MIT's CSAIL. T-Party focused on mobile information technologies and supported the work of 15 principal investigators and more than 50 PhD students. The additional funding will extend this work through 2015 and support new research into operating systems and programming environments for cloud computing. "Quanta is CSAIL's largest sponsor and our only strategic partner. For the past four years, we have jointly defined a multi-faceted research agenda around the common goal of future mobile communication. We look forward to working even closer with Quanta in the next six years on an equally ambitious goal of reengineering cloud computing," said CSAIL Director Victor Zue.Quanta Chairman and CEO Barry Lam said, "It is both our privilege and honor to engage with MIT CSAIL for another six years. We hope Quanta and MIT can jointly create innovative technologies and products both to enhance and enrich people's life. It is also our wish that we can contribute to the evolution of our culture through revolutionary engineering inventions."
Robotic therapy holds promise for cerebral palsy
Tue, 19 May 2009 05:00:00 -0000
Over the past few years, MIT engineers have successfully tested robotic devices to help stroke patients learn to control their arms and legs. Now, they’re building on that work to help children with brain injuries and disorders such as cerebral palsy."Robotic therapy can potentially help reduce impairment and facilitate neuro-development of youngsters with cerebral palsy," says Hermano Igo Krebs, principal research scientist in mechanical engineering and one of the project's leaders.Krebs and others at MIT, including professor of mechanical engineering Neville Hogan, pioneered the use of robotic therapy in the late 1980s, and since then the field has taken off. Above: Robotics: A New Hope in Cerebral Palsy. This video shows devices developed at MIT as well as non-MIT robots.View this post on MIT TechTV."We started with stroke because it's the biggest elephant in the room, and then started to build it out to other areas, including cerebral palsy as well as multiple sclerosis, Parkinson's disease and spinal cord injury," says Krebs.The team's suite of robots for shoulder-and-elbow, wrist, hand and ankle has been in clinical trials for more than 15 years with more than 400 stroke patients. The Department of Veterans Affairs has just completed a large-scale, randomized, multi-site clinical study with these devices. All the devices are based on the same principle: that it is possible to help rebuild brain connections using robotic devices that gently guide the limb as a patient tries to make a specific movement. When the researchers first decided to apply their work to children with cerebral palsy, Krebs was optimistic that it would succeed, because children's developing brains are more plastic than adults', meaning they are more able to establish new connections.The MIT team is focusing on improving cerebral palsy patients' ability to reach for and grasp objects. Patients handshake with the robot via a handle, which is connected to a computer monitor that displays tasks similar to those of simple video games.In a typical task, the youngster attempts to move the robot handle toward a moving or stationary target shown on the computer monitor. If the child starts moving in the wrong direction or does not move, the robotic arm gently nudges the child's arm in the right direction. Krebs began working in robotic therapy as a graduate student at MIT almost 20 years ago. In his early studies, he and his colleagues found that it's important for stroke patients to make a conscious effort during physical therapy. When signals from the brain are paired with assisted movement from the robot, it helps the brain form new connections that help it relearn to move the limb on its own.Even though a stroke kills many neurons, "the remaining neurons can very quickly establish new synapses or reinforce dormant synapses," says Krebs.For this type of therapy to be effective, many repetitions are required -- at least 400 in an hour-long session.Results from three published pilot studies involving 36 children suggest that cerebral palsy patients can also benefit from robotic therapy. The studies indicate that robot-mediated therapy helped the children reduce impairment and improve the smoothness and speed of their reaching motions.The researchers applied their work to stroke patients first because it is such a widespread problem -- about 800,000 people suffer strokes in the United States every year. About 10,000 babies develop cerebral palsy in the United States each year, but there is more potential for long-term benefit for children with cerebral palsy."In the long run, people that have a stroke, if they are 70 or 80 years old, might stay with us for an average of 5 or 6 years after the stroke," says Krebs. "In the case of cerebral palsy, there is a whole life."Most of the clinical work testing the device with cerebral palsy patients has been done at Blythedale Children's Hospital in Westchester County, N.Y., and Spaulding Rehabilitation Hospital in Boston. Other hospitals around the country and abroad are also testing various MIT-developed robotic therapy devices.Krebs' team has focused first on robotic devices to help cerebral palsy patients with upper body therapy, but they have also initiated a project to design a pediatric robot for the ankle. Among Krebs' and Hogan's collaborators on the cerebral palsy work are Dr. Mindy Aisen '76, former head of the Department of Veterans Affairs Office of Research and Development and presently the director and CEO of the Cerebral Palsy International Research Foundation (CPIRF); Dr. Joelle Mast, chief medical officer, and Barbara Ladenheim, director of research, of Blythedale Children's Hospital; and Fletcher McDowell, former CEO of the Burke Rehabilitation Hospital and a member of the CPIRF board of directors. MIT's work on robotic therapy devices is funded by CPIRF and the Niarchos Foundation, the Department of Veterans Affairs, the New York State NYSCORE, and the National Center for Medical Rehabilitation Research of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. A version of this article appeared in MIT Tech Talk on May 20, 2009 (download PDF).
'Bother bots' win the day
Fri, 08 May 2009 05:00:00 -0000
There were a variety of ways to score points in this year's 2.007 competition, which culminated in head-to-head (or wheel-to-wheel) matches among about 150 robots built by the students over the course of the semester. But one strategy seemed to prevail: preventing one's opponent from scoring, using a secondary "bother bot" to get in the way.The final contest, which had no effect on students' grades for the class but nevertheless spurred intense competition and effusive cheering, was held Thursday evening at the Johnson Athletic Center's ice rink. And the best bother bot brought home the gold -- or rather, its creator, sophomore Edward Grinnell, did.Asked to deliver a victory speech after the final round, Grinnell offered oration of machinelike economy and precision: "2.007 is awesome!" he said.2.007, which evolved from a class started in 1970 by Woodie Flowers SM '68, ME '71, PhD '73, the Pappalardo Professor of Mechanical Engineering, Emeritus, focuses on design and manufacturing and is a required class for sophomores in mechanical engineering. The class's traditional semester-ending competition features robots built mainly from identical kits of components issued to each student. The matches were played on a two-meter square playing field divided down the middle by a row of cinderblock "buildings" separated by alleys 3 inches wide. The robots, which operated autonomously for the first 10 seconds of each minute-long match and then were controlled using radio control devices, could score by moving blocks to a designated spot, extra points for stacking the blocks, more points for picking up crushed cans and placing them in a slot, and the highest scores for crushing a can and then placing it in the slot. The scores could also be multiplied by moving a boot, attached to a pulley, toward one's own side of the field -- something that none of the robots managed to do.Many students built elaborate can-crushing devices, some of which worked well in the preliminary elimination rounds on Wednesday. But because the bother bots were so effective in thwarting can crushers, not a single can was successfully crushed during the final contest, which featured the 32 highest-scoring bots."The bother bots seemed to rise to the top," said lead instructor Daniel Frey PhD '97, a professor of mechanical engineering and engineering systems. "A good defense often beats a good offense." Simple but robust strategies prevailed. In second place was a machine built by Pablo Bello, which also had a bother bot of its own but was defeated by Grinnell's more sturdy low-slung wedge-shaped bother bot. The third-place finisher, built by Elvine Pineda, was decorated with blue lights and was one of the most attractive robots in the contest; very effective in the early rounds, it quickly grabbed the pre-crushed cans and placed them in the slot. But in its semifinal matchup, it was successfully thwarted by Bello's bother bot, which prevented it from reaching the slot.Trophies and t-shirts were given to the top eight finishers, and the top four finishers will have an opportunity to attend a similar international robot design competition in Tokyo this summer. Organizers also presented the Whitelaw prize -- a special award for excellence in design and manufacturing -- to four competitors.Dick Fenner, director of the Pappalardo Lab, emphasized that while the competition is a fun and exciting conclusion for the class, just creating a novel design and building a machine that works at all, in the brief period of one semester, is a significant accomplishment. "If you put something on the table and it wiggles, you're a hero in my book," he said. A version of this article appeared in MIT Tech Talk on May 13, 2009 (download PDF).

Latest Issue of Journal of Experimental & Theoretical Artificial Intelligence

An empirical study of constraint logic programming and answer set programming solutions of combinatorial problems
Dovier, AgostinoFormisano, AndreaPontelli, Enrico
A model for the dynamic coordination of multiple competing goals
Martin H, Jose Antoniode Lope, Javier
Computational theories of mind, and Fodor's analysis of neural network behaviour
Guarini, Marcello

Subscribe to Artificial_Intelligence RSS feed

Singularity Institute for Artificial Intelligence