Reinforcement learning is a technique in which a computer system learns how best to solve some problem through trial-and-error. Classic applications of reinforcement learning involve problems like robot navigation and automated surveillance. Now, researchers have developed a new reinforcement-learning algorithm that, for many problems, allows computer systems to find solutions much more efficiently than previous algorithms did.
Researchers from North Carolina State University have developed a software algorithm that detects and isolates cyberattacks on networked control systems—which are used to coordinate transportation, power, and other infrastructure across the United States.
You get into your car and ask it to get you home in time for the start of the big game, stopping off at your favorite Chinese restaurant on the way for takeout. But the car informs you that the road past the Chinese restaurant is closed for repairs, and you will have to choose a different place. You select a nearby Korean restaurant from the options the car suggests. Autonomous devices could soon collaborate with humans in this way.
Whether reaching for a book out of a cluttered cabinet or pruning a bush in the backyard, a person’s arm frequently makes contact with objects during everyday tasks. Animals do it too, when foraging for food, for example. Much in the same way, robots are now able to intelligently maneuver within clutter, gently making contact with objects while accomplishing a task. This new control method has wide applications.
Surgical robots could make some types of surgery safer and more effective, but proving that the software controlling these machines works as intended is problematic. Researchers at Carnegie Mellon University and Johns Hopkins University have demonstrated that methods for reliably detecting software bugs and ultimately verifying software safety can be applied successfully to this breed of robot.
Researchers in the U.K. have been working to program a group of 40 robots to carry out simple fetching and carrying tasks, by grouping around an object and working together to push it across a surface. Even after being scattered, the robots can group again and organize themselves by order of priority. The team says the ability to control robot swarms could prove hugely beneficial in a range of contexts, from military to medical.
Robot butlers that tidy your house or cook you a meal have long been the dream of science-fiction writers and artificial intelligence researchers alike. But if robots are ever going to move effectively around our constantly changing homes or workspaces performing such complex tasks, they will need to be more aware of their own limitations, according to researchers at Massachusetts Institute of Technology.
When a robot is moving one of its limbs through free space, its behavior is well described by a few simple equations. But as soon as it strikes something solid, those equations break down. Roboticists typically use ad hoc control strategies to negotiate collisions and then revert to their rigorous mathematical models when the robot begins to move again. Researchers at Massachusetts Institute of Technology are hoping to change that, with a new mathematical framework that unifies the analysis of both collisions and movement through free space.
When Georgia Tech opens the doors to the Georgia Dome next month as the host institution for the 2013 Final Four, expect third-seeded Florida to walk out as the national champion. That's the prediction from Georgia Tech's Logistic Regression/Markov Chain (LRMC) college basketball ranking system, a computerized model that has chosen the men's basketball national champ in three of the last five years.
Memristors are made of fine nanolayers and can be used to connect electric circuits and for several years have been considered to be the electronic equivalent of the synapse. A researcher in Germany, physicist Andy Thomas, is now using his memristors as key components for his blueprint for an artificial brain.
Many commercial robotic arms perform what roboticists call "pick-and-place" tasks: The arm picks up an object in one location and places it in another. Usually, the objects are positioned so that the arm can easily grasp them; the appendage that does the grasping may even be tailored to the objects' shape. General-purpose household robots, however, would have to be able to manipulate objects of any shape, left in any location. And today, commercially available robots don't have anything like the dexterity of the human hand. Until now.
Ancient languages hold a treasure trove of information about the culture, politics and commerce of millennia past. Yet, reconstructing them to reveal clues into human history can require decades of painstaking work. Now, scientists at the University of California, Berkeley, have created an automated “time machine,” of sorts, that will greatly accelerate and improve the process of reconstructing hundreds of ancestral languages.
Research at the University of Gothenburg and Chalmers University of Technology has resulted in a new type of machine that sorts used batteries by means of artificial intelligence (AI). One machine is now being used in the U.K., sorting one-third of the country's recycled batteries.
Stanford University researchers have designed the fastest, most accurate algorithm yet for brain-implantable prosthetic systems that can help disabled people maneuver computer cursors with their thoughts. The algorithm's speed, accuracy, and natural movement approach those of a real arm, doubling performance of existing algorithms.
For anyone who has ever struggled while attempting to solve a Sudoku puzzle, University of Notre Dame researchers are riding to the rescue. They can not only explain why some Sudoku puzzles are harder than others, they have also developed a mathematical algorithm that solves Sudoku puzzles very quickly, without any guessing or backtracking.
One hundred years after the birth of mathematician and computer scientist Alan Turing, whose “Turing test” stands as one of the foundational definitions of what constitutes true machine intelligence, a virtual “gamer” created by computer scientists at The University of Texas at Austin has won the annual BotPrize by convincing a panel of judges that their software-based robot was more human-like than half the humans it competed against.
Developed by a company in San Diego, a new automated system that lets consumers trade in cell phones and mobile devices for reimbursement or recycling relies artificial intelliigence and sophisticated machine vision diagnostics. The building blocks for the ecoATM have existed for many years, but none, until now, have been applied to the particular problem of consumer recycling.
A new, Massachusetts Institute of Technology-developed analytical method identifies the precise binding sites of transcription factors—proteins that regulate the production of other proteins—with 10 times the accuracy of its predecessors.
Researchers in Malaysia have developed a system that allows a computer to “read lips”. The invention involves a genetic algorithm that gets better and better with each iteration to match irregular ellipse fitting equations to the shape of the human mouth displaying different emotions. The system could improve the way we interact with computers and perhaps allow disabled people to use communications devices more effectively.
Disorders such as schizophrenia can originate in certain regions of the brain and then spread out to affect connected areas. Identifying these regions of the brain, and how they affect the other areas they communicate with, would allow drug companies to develop better treatments and could ultimately help doctors make a diagnosis. But interpreting the vast amount of data produced by brain scans to identify these connecting regions has so far proved impossible, until now.
Most major Websites maintain huge databases. Almost any transaction on a shopping site, travel site, or social networking site require multiple database queries, which can slow response time. Now, researchers at Massachusetts Institute of Technology have developed a system that automatically streamlines Websites' database access patterns, making the sites up to three times as fast.
Research into the genetic factors behind certain disease mechanisms, illness progression, and response to new drugs is frequently carried out using tiny multicellular animals such as nematodes. Often progress relies on the microscopic visual examination of many individual animals to detect mutants worthy of further study. Now, scientists have demonstrated an automated system that uses artificial intelligence and image processing to examine large numbers of individual Caenorhabditis elegans .
Last spring private industry successfully sent a spacecraft carrying cargo to the International Space Station. Now the race is on to see which company will be the first to make commercial human spaceflight a reality. Sierra Nevada Corporation will receive hundreds of millions of dollars to further develop its commercial human spacecraft system, NASA announced earlier this month; and they are now turning to Georgia Tech for help.
Batoid rays, such as stingrays and manta rays, are among nature's most elegant swimmers. They are fast, highly maneuverable, graceful, energy efficient, can cruise, bird-like, for long distances in the deep, open ocean, and rest on the sea bottom. A team from the University of Virginia and other universities is trying to emulate the seemingly effortless, but powerful, swimming motions of rays by engineering their own ray-like machine modeled on nature.
Massachusetts Institute of Technology researchers have designed algorithms that vastly improve robots' navigation and feature-detecting capabilities. Using the group's algorithms, robots are able to swim around a ship's hull and view complex structures such as propellers and shafts. The goal is to achieve a resolution fine enough to detect a 10-cm mine attached to the side of a ship.