Syntactic and Semantic Filtering in a Chart Parser
نویسندگان
چکیده
This paper describes a method to enhance the performance of a unification-style bottom-up chart parser by means of top-down filtering techniques. The filter developed consists of a syntactic module which prevents the construction of redundant edges in the chart by ensuring that a proposed edge in the chart can really be syntactically combined with neighboring edges later, and a semantic module which ensures that the semantic information in a proposed edge in the chart is compatible with semantic information in other edges. 1 I n t r o d u c t i o n Chart parsers used in natural language processing parse an input sentence by building up a data structure called the chart. A chart is a network of vertices representing points in the sentence which are linked by edges which represent constituents of the sentence. In a unification-style chart parser [5] the chart is augmented step by step by adding pieces of description according to the grammar. Each partial description added remains in the chart and serves to constrain the possibilities for further augmentation. In LINK, a bottom-up unification-style chart parser described by Lytinen [3], a directed acyclic graph (DAG) is built to represent the analysis of a sentence. Edges built by LINK are labeled by DAGs incorporating syntactic as well as semantic information. New edges are added to the chart by applying unification rules. This guarantees that the new constituent added to the parse has the necessary syntactic and semantic features. We present a way of integrating bottom-up parsing with top-down parsing, both in terms of syntax and semantics. Our approach can be described as bottom-up parsing with top-down filtering. We modified LINK using this approach. Similar approaches have been variously called in the literature a "filter" [1] and an "oracle" [4]. 2 N e c e s s i t y F o r T o p d o w n F i l t e r i n g M e t h o d s In a top-down parser, parsing is rule-driven while in a bottom-up parser, parsing is data-driven. Thus in a top-down chart parser an active edge in the chart is sought to be expanded with all rules in the grammar which have the current symbol as the left-hand side, causing the parser to be over-productive in edge
منابع مشابه
برچسبزنی نقش معنایی جملات فارسی با رویکرد یادگیری مبتنی بر حافظه
Abstract Extracting semantic roles is one of the major steps in representing text meaning. It refers to finding the semantic relations between a predicate and syntactic constituents in a sentence. In this paper we present a semantic role labeling system for Persian, using memory-based learning model and standard features. Our proposed system implements a two-phase architecture to first identify...
متن کاملFeature Engineering in Persian Dependency Parser
Dependency parser is one of the most important fundamental tools in the natural language processing, which extracts structure of sentences and determines the relations between words based on the dependency grammar. The dependency parser is proper for free order languages, such as Persian. In this paper, data-driven dependency parser has been developed with the help of phrase-structure parser fo...
متن کاملبرچسبزنی خودکار نقشهای معنایی در جملات فارسی به کمک درختهای وابستگی
Automatic identification of words with semantic roles (such as Agent, Patient, Source, etc.) in sentences and attaching correct semantic roles to them, may lead to improvement in many natural language processing tasks including information extraction, question answering, text summarization and machine translation. Semantic role labeling systems usually take advantage of syntactic parsing and th...
متن کامل1 Are Efficient Natural Language Parsers Robust ?
This paper discusses the robustness of four efficient syntactic error-correcting parsing algorithms that are based on chart parsing with a context-free grammar. In this context, by robust we mean able to correct detectable syntactic errors. We implemented four versions of a bottom-up error-correcting chart parser: a basic bottom-up chart parser, and chart parsers employing selectivity, top-down...
متن کاملSyntactic Recovery and Spelling Correction of Ill-formed Sentences
This paper describes syntactic repair and spelling correction of ill-formed sentences within a context-free grammar using non-static filtering, of ill-formed sentences which violate subjectverb agreement or premodifier-noun agreement. The system described here provides recovery of local trees, reconstruction of the sentence, and spelling correction of detected typographical errors. It also prod...
متن کامل