Group items tagged

"Falsehood flies, and the Truth comes limping after it," Jonathan Swift once wrote. It was hyperbole three centuries ago. But it is a factual description of social media, according to an ambitious and first-of-its-kind study produced by MIT and published Thursday in Science. The massive new study analyzes every major contested news story in English across the span of Twitter's existence-some 126,000 stories, tweeted by 3 million users, over more than 10 years-and finds that the truth simply cannot compete with hoax and rumor. By every common metric, falsehood consistently dominates the truth on Twitter, the study finds: Fake news and false rumors reach more people, penetrate deeper into the social network, and spread much faster than accurate stories. "It seems to be pretty clear [from our study] that false information outperforms true information," said Soroush Vosoughi, a data scientist at MIT who has studied fake news since 2013 and who led this study. "And that is not just because of bots. It might have something to do with human nature."

Two University of Washington undergraduates have won a $10,000 Lemelson-MIT Student Prize for gloves that can translate sign language into text or speech. The Lemelson-MIT Student Prize is a nationwide search for the most inventive undergraduate and graduate students. This year, UW sophomores Navid Azodi and Thomas Pryor - who are studying business administration and aeronautics and astronautics engineering, respectively - won the "Use It" undergraduate category that recognizes technology-based inventions to improve consumer devices. Their invention, "SignAloud," is a pair of gloves that can recognize hand gestures that correspond to words and phrases in American Sign Language. Each glove contains sensors that record hand position and movement and send data wirelessly via Bluetooth to a central computer. The computer looks at the gesture data through various sequential statistical regressions, similar to a neural network. If the data match a gesture, then the associated word or phrase is spoken through a speaker.

When your brain encounters sensory stimuli, such as the scent of your morning coffee or the sound of a honking car, that input gets shuttled to the appropriate brain region for analysis. The coffee aroma goes to the olfactory cortex, while sounds are processed in the auditory cortex. That division of labor suggests that the brain's structure follows a predetermined, genetic blueprint. However, evidence is mounting that brain regions can take over functions they were not genetically destined to perform. In a landmark 1996 study of people blinded early in life, neuroscientists showed that the visual cortex could participate in a nonvisual function - reading Braille. Now, a study from MIT neuroscientists shows that in individuals born blind, parts of the visual cortex are recruited for language processing. The finding suggests that the visual cortex can dramatically change its function - from visual processing to language - and it also appears to overturn the idea that language processing can only occur in highly specialized brain regions that are genetically programmed for language tasks. "Your brain is not a prepackaged kind of thing. It doesn't develop along a fixed trajectory, rather, it's a self-building toolkit. The building process is profoundly influenced by the experiences you have during your development," says Marina Bedny, an MIT postdoctoral associate in the Department of Brain and Cognitive Sciences and lead author of the study, which appears in the Proceedings of the National Academy of Sciences the week of Feb. 28.

This article talks about a study that was conducted by MIT linguistics professors and a group from Carleton University that explored the phenomenon of why children mix up the words "and" and "or." Linguists say that children use almost entirely the same approach as adults when it comes to evaluating potentially ambiguous sentences, by testing and "strengthening" them into sentences with more precise meanings, when disjunction and conjunction ("or" and "and") are involved. However, they found that children do not test how a sentence would change if "and" was directly substituted for "or." On the other hand, adults compute "scalar implicatures," which is a technical phrase for thinking about the implications of the logical relationship between a sentence and its alternative. The research team conducted the study's experiment by testing 59 English-speaking children and 26 adults. The children ranged in age from 4 months to 6 years. The linguists gave the subjects a series of statements along with pictures, and asked them to say whether the statements were true or false. The results suggest that children are computing scalar implicatures when they evaluate the statements, but they largely do not substitute disjunctions and conjunctions when testing out the possible meaning of sentences, as adults do. In general, the researchers observed, across languages, and for children and adults alike, when you take 'and' out of the space of alternatives, "or" becomes "and."

A new study of 37 languages by three MIT researchers has shown that most languages move toward "dependency length minimization" (DLM) in practice. That means language users have a global preference for more locally grouped dependent words, whenever possible. Apparently, it's easier for our brains to process a sentence when related words are closer together.

Here's a quick task: Take a look at the sentences below and decide which is the most effective. (1) "John threw out the old trash sitting in the kitchen." (2) "John threw the old trash sitting in the kitchen out." Either sentence is grammatically acceptable, but you probably found the first one to be more natural. A new study of 37 languages by three MIT researchers has shown that most languages move toward "dependency length minimization" (DLM) in practice. That means language users have a global preference for more locally grouped dependent words, whenever possible.

Language takes an astonishing variety of forms across the world-to such a huge extent that a long-standing debate rages around the question of whether all languages have even a single property in common. Well, there's a new candidate for the elusive title of "language universal" according to a paper in this week's issue of PNAS.

"In one 40-second clip, you can hear how "gaga" turned into "water" over the course of six months. In a video clip, below, you can hear and watch the evolution of "ball." .... Unreal 3-D visualizations allowed his team to zoom through the house like a dollhouse and map the utterance of each word in its context. In a landscape-like image with peaks and valleys, you can see that the word "water" was uttered most often in the kitchen, while "bye" took place at the door."

textbook

This article talks about how even though learning how to program a computer seems like learning a new language, computer language is actually processed through a different area of the brain than when language is processed. It says that it activates the multiple demand network, which is also activated by complex cognitive tasks such as solving math problems or crossword puzzles. However, it appears to rely on different parts of the network than math or logic problems do, which suggests that coding does not precisely replicate the cognitive demands of mathematics either. The article says that researchers didn't find any regions that appear to be exclusively devoted to programming. They also suggest that understanding computer code cannot be categorized as a language-based or math-based skill. Instead, they think it might require a combination of both skills, even though the brain regions don't reflect it.

This MIT Technology Review article notes that while Artificial Intelligence has experienced many sophisticated advances, one fundamental capability remains elusive: language. Systems like Siri and IBM's Watson can follow simple spoken or typed commands and answer basic questions, but they can't hold a conversation and have no real understanding of the words they use. In addition, humans, unlike machines, have the ability to learn very quickly from a relatively small amount of data and have a built-in ability to model the world in 3-D very efficiently. Programming machines to comprehend and generate language is a complex task, because the machines would need to mimic human learning, mental model building, and psychology. As MIT cognitive scientist Josh Tenenbaum states, "Language builds on other abilities that are probably more basic, that are present in young infants before they have language: perceiving the world visually, acting on our motor systems, understanding the physics of the world or other agents' goals." ­ If he is right, then it will be difficult to re-create language understanding in machines and AI systems without trying to mimic human learning, mental model building, and psychology.

A study done based on 670,000 people's grammar quiz results.

Bradley Hayes, a post-doc student at the Massachusetts Institute of Technology, has invented @DeepDrumpf, an amusing bit of artificial intelligence. DeepDrumpf is a bot trained on the publicly available speech transcripts, tweets, and debate remarks of Donald Trump. Using a machine learning model known as a Recurrent Neural Network, the bot generates sequences of words based on priming text and the statistical structure found within its training data. Created to highlight the absurdity of this election cycle, it has amassed over 20,000 followers and has been viewed over 12 million times -- showcasing the consequences of training a machine learning model on a dataset that embodies fearmongering, bigotry, xenophobia, and hypernationalism. Here's a sample tweet: "We have to end education. What they do is unbelievable, how bad. Nobody can do that like me. Believe me."

This NYT article reports findings of a recently released study, which concluded that computers are capable of scoring essays on standardized tests as well, or possibly better, than humans. However, Les Perelman, a director of writing at MIT and a tester of the Educational Testing Services E-Rater program disagrees, arguing that the system can be easily "gamed".

An interesting story--my students, you might recall Sheryl Turkle of MIT referencing robot therapists in her TED talk. USC has developed a robot therapist, Ellie, designed to talk to people who are struggling emotionally, and to take their measure in a way no human can. Originally developed to work with military PTSD patients, Ellie's purpose: to gather information and provide real human therapists detailed analysis of patients' movements and vocal features, in order to give new insights into people struggling with emotional issues. The body, face and voice express things that words sometimes obscure. Ellie's makers believe that her ability to do this will ultimately revolutionize American mental health care.

includes a section about song lyrics and a good introduction to the cognitive aspects of memory Excerpt: "The experimental data corroborate our intuition that the memory representation for lyrics seems to be tied into the memory representation for melody (Serafine, Crowder, and Repp 1984). Further evidence of this comes from a case report of a musician who suffered a stroke caused by blockage of the right cerebral artery. After the stroke, he was able to recognize songs played on the piano if they were associated with words (even though the words weren't being presented to him), but he was unable to recognize songs that were purely instrumentals (Steinke, Cuddy, and Jacobson 1995)."