MIT researchers have developed a database of annotated English words written by non-native English speakers, reports Kevin Hartnett for The Boston Globe. The database will provide “a platform for the study of learner English and also make it easier to develop technology like better search engines that supports non-native speakers.”
In an article for The Huffington Post about why virtual assistants have trouble understanding accents, Philip Ellis highlights how researchers from MIT have compiled a database of written English composed by non-native speakers. Ellis explains that the aim is "to create a richer context for machine learning” systems.
CSAIL researchers recently presented an algorithm that teaches computers to predict sounds, writes Kevin Hartnett for The Boston Globe. The ability to predict sounds will help robots successfully navigate the world and “make sense of what’s in front of them and figure out how to proceed,” writes Hartnett.
Associate Prof. Scott Aaronson answers the question “Is machine learning currently overhyped?” for Forbes. “I suppose it’s less interesting to me to look at the sheer amount of machine learning hype than at its content. Almost everyone in the 1950s knew that computers were going to be important, but they were often wildly wrong about the reasons,” he writes.
By watching TV shows and video clips, CSAIL researchers show that artificially intelligent systems can learn and predict human behavior, writes Tim Moynihan for Wired. Researchers say these findings could lead to analyzing hospital video feeds to alert emergency responders or allow robots to respond.
CSAIL researchers used videos of popular TV shows to train an algorithm to predict how two people will greet one another. “[T]he algorithm got it right more than 43 percent of the time, as compared to the shoddier 36 percent accuracy achieved by algorithms without the TV training,” notes Janet Burns in Forbes.
Mary Beth Griggs writes for Popular Science that CSAIL researchers have created an algorithm that can predict human interaction. Griggs explains that the algorithm could “lead to artificial intelligence that is better able to react to humans or even security cameras that could alert authorities when people are in need of help.”
Dan Misener writes for CBC News that CSAIL researchers have developed an algorithm that can predict interactions between two people. PhD student Carl Vondrick explains that the algorithm is "learning, for example, that when someone's hand is outstretched, that means a handshake is going to come."
CSAIL researchers have trained a deep-learning program to predict interactions between two people, writes Hope King for CNN. “Ultimately, MIT's research could help develop robots for emergency response, helping the robot assess a person's actions to determine if they are injured or in danger,” King explains.
In an article for Wired, Tim Moynihan writes that a team of CSAIL researchers has created a machine-learning system that can produce sound effects for silent videos. The researchers hope that the system could be used to “help robots identify the materials and physical properties of an object by analyzing the sounds it makes.”
Writing for the Financial Times, Clive Cookson reports that MIT researchers have developed an artificial intelligence system capable of producing realistic sounds for silent movies. Cookson explains that another application for the system could be “to help robots understand objects’ physical properties and interact better with their surroundings."
Washington Post reporter Matt McFarland writes that MIT researchers have created an algorithm that can produce realistic sounds. “The findings are an example of the power of deep learning,” explains McFarland. “With deep learning, a computer system learns to recognize patterns in huge piles of data and applies what it learns in useful ways.”
Popular Science reporter Mary Beth Griggs writes that MIT researchers have developed an algorithm that can learn how to predict sound. The algorithm “can watch a silent movie and create sounds that go along with the motions on screen. It's so good, it even fooled people into thinking they were actual, recorded sounds from the environment.”
Cade Metz writes for Wired that MIT researchers have developed a system that allows robots to predict how objects will move. Postdoc Ilker Yildirim explains that in order for a robot to be able to assist with household tasks like washing the dishes, it must “deeply understand its physical environments.”
MIT computer scientists have developed a program that can predict how objects will move with the same accuracy as humans, reports Mary Beth Griggs for Popular Science. The researchers hope to eventually be able to program the system to “make predictions in the natural world even faster than we can.”