FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W96702405> ?p ?o ?g. }

• W96702405 abstract "We present a bottom-up algorithm for parsing and generation. The algorithm is a bottom-up chart parser, whose lexical lookup phase has been modified for generation. An analysis of the algorithm offers interesting insights into the relationship between parsing and generation, summarized by the statement that parsing is a very constrained form of generation. The use of the generation algorithm as a component of a grammar development environment is discussed. 1 Parsing and Generation In this paper, we present a chart-based algorithm for parsing and generation. We will not consider generation from a semantic representation, for which the usual head-driven algorithms [Shieber et al. 1990] are better suited, but rather the generation of large sets of sentences. This functionality is needed in grammar development systems for exploring the coverage of a given grammar, and for checking whether a grammar overgenerates, which is particularly useful if grammars are developed not only for analysis, but also for generation. Let us start by giving an intuitive characterization of bottom-up parsing and generation. * This work was conducted in IBM Germany's text understanding project LILOG. I would like to thank Roman Georg Arens, Monika Becker, Gunter Neumann, Karel Oliva and Hans Uszkoreit for fruitful discussion. Bottom-up parsing starts by creating an item for every occurence of a word in the input string. Items carry information about the portion of the string they cover, their position in the string, and their syntactic category. If adjacent items match the right-hand side of a grammar rule, they are combined into larger items until an item is found that covers the entire string and whose category is whatever the grammar defines as a well-formed structure. The parsing process produces as many solutions as there are different derivations for the given string. In bottom-up generation, one initial item is created for every entry of the lexicon. In the case of generation, items need not carry information about string position. Any sequence of items whose categories match the right-hand side of a grammar rule can be combined into a larger item whose category is the left-hand side of the grammar rule. This process continues until no more items can be built. As any interesting language contains an infinite number of sentences, such an algorithm never stops generating items unless it is somehow restricted (see section 3). Under this view, parsing and generation differ in two respects: 1. In the initialization phase, a parser uses only the words of the string, whereas a generator uses all words of the lexicon. 2. Items used in parsing are indexed for their position in the string in order to check adjacency, whereas items used for generation carry information about the string they cover, but not about string position.These differences are illustrated in figure 1. These differences suggest that parsing is a very restricted form of generation, in which it is already known which sentence must be generated. A similar point on the relationship between parsing and generation is made in [Kay 1980]." @default.
• W96702405 created "2016-06-24" @default.
• W96702405 creator A5011263407 @default.
• W96702405 date "2000-01-01" @default.
• W96702405 modified "2023-09-27" @default.
• W96702405 title "A Bottom-Up Algorithm for Parsing and Generation" @default.
• W96702405 cites W1511588645 @default.
• W96702405 cites W1577398775 @default.
• W96702405 cites W1792465007 @default.
• W96702405 cites W2036531226 @default.
• W96702405 cites W2143745167 @default.
• W96702405 hasPublicationYear "2000" @default.
• W96702405 type Work @default.
• W96702405 sameAs 96702405 @default.
• W96702405 citedByCount "2" @default.
• W96702405 crossrefType "journal-article" @default.
• W96702405 hasAuthorship W96702405A5011263407 @default.
• W96702405 hasConcept C11413529 @default.
• W96702405 hasConcept C118364021 @default.
• W96702405 hasConcept C138885662 @default.
• W96702405 hasConcept C147547768 @default.
• W96702405 hasConcept C154945302 @default.
• W96702405 hasConcept C157486923 @default.
• W96702405 hasConcept C186644900 @default.
• W96702405 hasConcept C199360897 @default.
• W96702405 hasConcept C204321447 @default.
• W96702405 hasConcept C26022165 @default.
• W96702405 hasConcept C33923547 @default.
• W96702405 hasConcept C37914503 @default.
• W96702405 hasConcept C39890363 @default.
• W96702405 hasConcept C41008148 @default.
• W96702405 hasConcept C41895202 @default.
• W96702405 hasConcept C42560504 @default.
• W96702405 hasConcept C53893814 @default.
• W96702405 hasConcept C60690694 @default.
• W96702405 hasConceptScore W96702405C11413529 @default.
• W96702405 hasConceptScore W96702405C118364021 @default.
• W96702405 hasConceptScore W96702405C138885662 @default.
• W96702405 hasConceptScore W96702405C147547768 @default.
• W96702405 hasConceptScore W96702405C154945302 @default.
• W96702405 hasConceptScore W96702405C157486923 @default.
• W96702405 hasConceptScore W96702405C186644900 @default.
• W96702405 hasConceptScore W96702405C199360897 @default.
• W96702405 hasConceptScore W96702405C204321447 @default.
• W96702405 hasConceptScore W96702405C26022165 @default.
• W96702405 hasConceptScore W96702405C33923547 @default.
• W96702405 hasConceptScore W96702405C37914503 @default.
• W96702405 hasConceptScore W96702405C39890363 @default.
• W96702405 hasConceptScore W96702405C41008148 @default.
• W96702405 hasConceptScore W96702405C41895202 @default.
• W96702405 hasConceptScore W96702405C42560504 @default.
• W96702405 hasConceptScore W96702405C53893814 @default.
• W96702405 hasConceptScore W96702405C60690694 @default.
• W96702405 hasLocation W967024051 @default.
• W96702405 hasOpenAccess W96702405 @default.
• W96702405 hasPrimaryLocation W967024051 @default.
• W96702405 hasRelatedWork W1595695567 @default.
• W96702405 hasRelatedWork W2011544166 @default.
• W96702405 hasRelatedWork W2090956276 @default.
• W96702405 hasRelatedWork W2117347528 @default.
• W96702405 hasRelatedWork W2136473550 @default.
• W96702405 hasRelatedWork W2164854335 @default.
• W96702405 hasRelatedWork W2179349885 @default.
• W96702405 hasRelatedWork W2281356076 @default.
• W96702405 hasRelatedWork W2294746367 @default.
• W96702405 hasRelatedWork W2561271798 @default.
• W96702405 hasRelatedWork W2611033543 @default.
• W96702405 hasRelatedWork W2934008215 @default.
• W96702405 hasRelatedWork W2949296763 @default.
• W96702405 hasRelatedWork W3037124152 @default.
• W96702405 hasRelatedWork W3037405950 @default.
• W96702405 hasRelatedWork W3088554114 @default.
• W96702405 hasRelatedWork W327314173 @default.
• W96702405 hasRelatedWork W1847655440 @default.
• W96702405 hasRelatedWork W2124225806 @default.
• W96702405 hasRelatedWork W2181868809 @default.
• W96702405 isParatext "false" @default.
• W96702405 isRetracted "false" @default.
• W96702405 magId "96702405" @default.
• W96702405 workType "article" @default.