Powerset's superiority lies in the three decades of hard work by scientists at PARC. (PARC licensed much of its natural-language search technology to Powerset in February.) There was not one piece of technology that solved the problem, Pell says, but instead, it was the unification of many theories and fragments that pulled the project together."After 30 years, it's finally reached a point where it can be brought into the world," he says.A key component of the search engine is a deep natural-language processing system that extracts the relationships between words; the system was developed from PARC's Xerox Linguistic Environment (XLE) platform.
In the last 10 years, it's become far more common for physicians to keep records electronically. Those records could contain a wealth of medically useful data: hidden correlations between symptoms, treatments and outcomes, for instance, or indications that patients are promising candidates for trials of new drugs. Much of that data, however, is buried in physicians’ freeform notes.
A new Cornell study suggests language use is simpler than they had thought.
Co-author Morten Christiansen, Cornell professor of psychology and co-director of the Cornell Cognitive Science Program, and his colleagues say that language is actually based on simpler sequential structures, like clusters of beads on a string.
"What we're suggesting is that the language system deals with words by grouping them into little clumps that are then associated with meaning," he said. ...
For more than 50 years, language scientists have assumed that sentence structure is fundamentally hierarchical, made up of small parts in turn made of smaller parts, like Russian nesting dolls. language is naturally sequential, given the temporal cues that help us understand and be understood as we use language.
Tags Google began rolling out a feature that gives searchers in the United States the potential to access more relevant and in-depth responses to answers without leaving the page. The search results page displays a variety of content related to keyword queries, bringing up a list of facts, photos, and landmarks, as well as quick links to other popular uses for the search term. Rob Garner, vice president of strategy at agency iCrossing, said Google's knowledge graph takes another step in the company's long transition to develop an artificial intelligence engine -- semantic search. The change represents an effort by search engines to move away from text-based links in search results and serve up knowledge in fewer clicks.
Despite Watson's tremendous performance, the Final Jeopardy question at the end of Tuesday night's airing revealed the Achilles' heel that computer scientists have known all along: Watson doesn't really "think" anything, and it struggles with simple questions that most humans can answer without a second thought. Most of the clues on the "Jeopardy" board mention proper nouns -- specific places, events, people, songs, books and so on, says Dr. Douglas Lenat, a machine learning pioneer, former Stanford professor of computer science and CEO of Cycorp, a company that develops semantic technologies. But if that date had been part of the clue, could Watson correctly pick out [Schubert's] maternal grandmother's birth date from a list where only one of the dates was earlier than 1797?" We could, because we understand that everyone is younger than their own mother and grandmother, but Watson is unable to understand this, Lenat explained.
This thesis describes a formal model of a subtype of humour, and the implementation of that model in a program that generates jokes of that subtype. Although there is a great deal of literature on humour in general, very little formal work has been done on puns, and none has been implemented. All current linguistic theories of humour are over-general and not falsifiable. Our model, which is specific, formal, implemented and evaluated, makes a significant contribution to the field. Punning riddles are our chosen subtype of verbal humour, for several reasons. They are very common, they exhibit certain regular structures and mechanisms, and they have been studied previously by linguists. Our model is based on our extensive analysis of large numbers of punning riddles, taken from children's joke books. The implementation of the model, JAPE (Joke Analysis and Production Engine), generates punning riddles, from a humour independent lexicon. Pun generation requires much less world knowledge than pun comprehension, making it feasible for implementation. To support our claim that all of JAPE's output is punning riddles, we conducted an evaluatory experiment. We took JAPE texts, human-generated texts, nonsense non-jokes and sensible non-jokes, and asked joke experts to evaluate them. For joke experts, we used 8-11 year old children, since psychological research suggests that this age group enjoys, and can recognize, punning riddles better than other age groups. The results showed that JAPE's output texts are, in fact, recognizably jokes. The evaluation showed that our model adequately describes a significant subtype of verbal humour. We believe that this model can now be expanded to cover puns in general, as well as other types of linguistic humour.
Can computers learn to read? We think so. "Read the Web" is a research project that attempts to create a computer system that learns over time to read the web. Since January 2010, our computer system called NELL (Never-Ending Language Learner) has been running continuously...
WordNet® is a large lexical database of English. Nouns, verbs, adjectives and adverbs are grouped into sets of cognitive synonyms (synsets), each expressing a distinct concept. Synsets are interlinked by means of conceptual-semantic and lexical relations. The resulting network of meaningfully related words and concepts can be navigated with the browser. WordNet is also freely and publicly available for download. WordNet's structure makes it a useful tool for computational linguistics and natural language processing.
The leading voice-recognition company in China could use its momentum to expand internationally. You might not have heard of iFlyTek.
What would it take to develop machine learners that run forever, each day improving their performance and also the accuracy with which they learn? This talk will describe our attempt to build a never-ending language learner, NELL, that runs 24 hours per day, forever, and that each day has two goals: (1) extract more structured information from the web to populate its growing knowledge base, and (2) learn to read better than yesterday, by using previously acquired knowledge to better constrain its subsequent learning.The approach implemented by NELL is based on two key ideas: coupling the semi-supervised training of hundreds of different functions that extract different types of information from different web sources, and automatically discovering new constraints that more tightly couple the training of these functions over time. NELL has been running nonstop since January 2010 (follow it at CMU Read the Web project) and had extracted a knowledge base containing hundreds of thousands of beliefs as of June 2010. This talk will describe NELL, its successes and its failures, and use it as a case study to explore the question of how to design never-ending learners.
What would it take to develop machine learners that run forever, each day improving their performance and also the accuracy with which they learn? This talk will describe our attempt to build a never-ending language learner, NELL, that runs 24 hours per day, forever, and that each day has two goals: (1) extract more structured information from the web to populate its growing knowledge base, and (2) learn to read better than yesterday, by using previously acquired knowledge to better constrain its subsequent learning.
The approach implemented by NELL is based on two key ideas: coupling the semi-supervised training of hundreds of different functions that extract different types of information from different web sources, and automatically discovering new constraints that more tightly couple the training of these functions over time. NELL has been running nonstop since January 2010 (follow it at CMU Read the Web project) and had extracted a knowledge base containing hundreds of thousands of beliefs as of June 2010. This talk will describe NELL, its successes and its failures, and use it as a case study to explore the question of how to design never-ending learners.
Includes MP3 audio and MP4, Ogg Theora, and Windows Media video downloads.
In order to answer questions about children's stories one needs a great deal of "common sense" knowledge. A model Is presented which gives a rough organization to this knowledge along with specifications as to how the Information will be accessed. This rough model Is then used as a basis for tight arguments about narrow Issues (primarily using examples concerning piggy banks.) The paper Is intended as an illustration of how one might go about constructing a theory of knowledge.
The parser described in this paper is a conceptual parser. It s primary concern is to explicate the underlying meaning and conceptual relationships present in a piece of discourse in any natural language. It s output is a language-free network consisting of unambiguous concepts and their relations to other concepts. Pieces of discourse with identical meanings, whether in the same or different languages, parse into the same conceptual network.
A program was writte n to solve calculus word problems. The program, CARPS (CAlculus Rate Problem Solver), is restricte d to rate problems. The overall plan of the program is simila r to Bobrow's STUDENT, the primary difference being the introductio n of "structures " as the internal model in CARPS. Structures are stored internally as trees, each structure holding the information gathered about one object.
... we began to program a computer understanding system thatwould attempt to process input texts. An item crucial to our ability to accomplishthis task was what we called a script. A script is a frequently repeated causalchain of events that describes a standard situation. In understanding, when it ispossible to notice that one of these standard event chains has been initiated,then it is possible to understand predictively. That is, if we know we are in arestaurant then we can understand where an "order" fits with what we justheard, who might be ordering what from whom, what preconditions (menu,sitting down) might have preceded the "order", and what is likely to happennext. All this information comes from the restaurant script.