|
Language is so much more than a string of words. To understand whatsomeone means, you need context. Consider the phrase, "Man onfirst." It doesn't make much sense unless you're at a baseballgame. Or imagine a sign outside a children's boutique that reads,"Baby sale - One week only!" You easily infer from the situationthat the store isn't selling babies but advertising bargains ongear for them. Present these widely quoted scenarios to a computer, however, andthere would likely be a communication breakdown. Computers aren'tvery good at pragmatics - how language is used in socialsituations. But a pair of Stanford psychologists has taken the first stepstoward changing that. In a new paper published recently in the journal Science, AssistantProfessors Michael Frank and Noah Goodman describe a quantitativetheory of pragmatics that promises to help open the door to morehuman-like computer systems, ones that use language as flexibly aswe do. The mathematical model they created helps predict pragmaticreasoning and may eventually lead to the manufacture of machinesthat can better understand inference, context and social rules. Thework could help researchers understand language better and treatpeople with language disorders. It also could make speaking to a computerized customer serviceattendant a little less frustrating. "If you've ever called an airline, you know the computer voicerecognizes words but it doesn't necessarily understand what youmean," Frank said. "That's the key feature of human language. Insome sense it's all about what the other person is trying to tellyou, not what they're actually saying." Frank and Goodman's work is part of a broader trend to try tounderstand language using mathematical tools. That trend has led totechnologies like Siri, the iPhone's speech recognition personalassistant. But turning speech and language into numbers has its obstacles,mainly the difficulty of formalizing notions such as "commonknowledge" or "informativeness." That is what Frank and Goodman sought to address. The researchers enlisted 745 participants to take part in an onlineexperiment. The participants saw a set of objects and were asked tobet which one was being referred to by a particular word. For example, one group of participants saw a blue square, a bluecircle and a red square. The question for that group was: Imagineyou are talking to someone and you want to refer to the middleobject. Which word would you use, "blue" or "circle"? The other group was asked: Imagine someone is talking to you anduses the word "blue" to refer to one of these objects. Which objectare they talking about? "We modeled how a listener understands a speaker and how a speakerdecides what to say," Goodman explained. The results allowed Frank and Goodman to create a mathematicalequation to predict human behavior and determine the likelihood ofreferring to a particular object. "Before, you couldn't take these informal theories of linguisticsand put them into a computer. Now we're starting to be able to dothat," Goodman said. The researchers are already applying the model to studies onhyperbole, sarcasm and other aspects of language. "It will take years of work but the dream is of a computer thatreally is thinking about what you want and what you mean ratherthan just what you said," Frank said. We are high quality suppliers, our products such as Coil Winding Nozzle Manufacturer , China Yokogawa Tension Meter for oversee buyer. To know more, please visits Mechanical Wire Tensioner.
Related Articles -
Coil Winding Nozzle Manufacturer, China Yokogawa Tension Meter,
|
