We present a method for extracting parts of objects from wholes (e.g. "speedometer" from "car"). Given a very large corpus our method finds part words with 55% accuracy for the top 50 words as ranked by the system. The part list could be scanned by an end-user and added to an existing ontology (such as WordNet), or used as a part of a rough semantic lexicon.

ConceptNet is a freely available commonsense knowledgebase and natural-language-processing toolkit which supports many practical textual-reasoning tasks over real-world documents including topic-jisting (e.g. a news article containing the concepts, "gun," "convenience store," "demand money" and "make getaway" might suggest the topics "robbery" and "crime"), affect-sensing (e.g. this email is sad and angry), analogy-making (e.g. "scissors," "razor," "nail clipper," and "sword" are perhaps like a "knife" because they are all "sharp," and can be used to "cut something"), and other contextoriented inferences. The knowledgebase is a semantic network presently consisting of over 1.6 million assertions of commonsense knowledge encompassing the spatial, physical, social, temporal, and psychological aspects of everyday life. Whereas similar large-scale semantic knowledgebases like Cyc and WordNet are carefully handcrafted, ConceptNet is generated automatically from the 700,000 sentences of the Open Mind Common Sense Project -- a World Wide Web based collaboration with over 14,000 authors.

WordNet is a rich source of world knowledge from which formal axioms can be derived. In this paper we present a method for transforming the WordNet glosses into logic forms and further into axioms. The transformation of WordNet glosses into logic forms is useful for theorem proving and other applications. The paper demonstrates the utility of WordNet axioms in a question answering system to rank and extract answers.

As one attack on the "knowledge acquisition bottleneck", we are attempting to exploit a largely untapped source of general knowledge in texts, lying at a level beneath the explicit assertional content. This knowledge consists of relationships implied to be possible in the world, or, under certain conditions, implied to be normal or commonplace in the world. The goal of the work reported is to derive such general world knowledge (initially, from Penn Treebank corpora) in two stages: first, we derive general "possibilistic" propositions from noun phrases and clauses; then we try to derive stronger generalizations, based on the nature and statistical distribution of the possibilistic claims obtained in the first phase. Here we report preliminary results of the first phase, which indicate the feasibility of our project, and its likely limitations.

Abstract not found

Word sense disambiguation (WSD) is the ability to identify the meaning of words in context in a computational manner. WSD is considered an AI-complete problem, that is, a task whose solution is at least as hard as the most difficult problems in artificial intelligence. We introduce the reader to the motivations for solving the ambiguity of words and provide a description of the task. We overview supervised, unsupervised, and knowledge-based approaches. The assessment of WSD systems is discussed in the context of the Senseval/Semeval campaigns, aiming at the objective evaluation of systems participating in several different disambiguation tasks. Finally, applications, open problems, and future directions are discussed.

Abstract. This paper describes an automatic approach to identify lexical patterns which represent semantic relationships between concepts, from an on-line encyclopedia. Next, these patterns can be applied to extend existing ontologies or semantic networks with new relations. The experiments have been performed with the Simple English Wikipedia and WordNet 1.7. A new algorithm has been devised for automatically generalising the lexical patterns found in the encyclopedia entries. We have found general patterns for the hyperonymy, hyponymy, holonymy and meronymy relations and, using them, we have extracted more than 1200 new relationships that did not appear in WordNet originally. The precision of these relationships ranges between 0.61 and 0.69, depending on the relation. 1

ly have a separate entry for each category; for example, flap would have one entry as a noun and another as a verb. Separate entries are usually also appropriate for each of the senses of a homonym---a word that has more than one unrelated sense even within a single syntactic category; for example, the noun pen would have distinct entries for the senses writing instrument, animal enclosure,andswan. Polysemy--- related or overlapping senses---is a more-complex situation; sometimes the senses may be discrete enough that we can treat them as distinct: for example, window as both opening in wall and glass pane in opening in wall (fall through the window; break the window). But this is not always so; the word open, for example, has many overlapping senses concerning unfolding, expanding, revealing, moving to an open position, making openings in, and so on, and separating them into discrete senses, as the writers of dictionary definitions try to do, is not possible (see also sections 1.2.3 a

The diversity and availability of information sources on the World Wide Web has set the stage for integration and reuse at an unparalleled scale. There remain significant hurdles to exploiting the extent of the Web's resources in a consistent, scalable and maintainable fashion. The autonomy and volatility of Web sources complicates maintaining wrappers consistent with the requirements of the data's target application. This paper describes the ArcRank model of relationships between nodes in a directed labeled graph, such as hypertext. The paper presents a ranking algorithm for directed arcs, and the algorithm for extraction of hierarchical relationships between words in a dictionary. Using ArcRank we create a thesaurus style tool to aid in the integration of texts and databases whose content is similar but whose terms are different. These algorithms complement handcrafted thesauri, by determining more complete relationships between words, although they are less specific. Exploiting hier...

Although constructed for researchers to share news and information, Usenet quickly developed into a social environment with varied styles of interactions. Unfortunately, the browsers developed to view the shared messages fail to effectively convey the rich social features of a newsgroup, let alone all of Usenet. The goal of our research is to use the salient features of social interaction to build a “legible ” interactive visual representation of Usenet. In this paper, we introduce our approach to developing this type of visualization, discussing our theoretical framework, questions considered to access the socially salient features, and a series of design iterations used for exploring how to develop a visual language that conveys social meaning. Although this paper represents a work in progress, we hope that this approach and our initial iterations help build a framework for future directions.