Bangalore Srinivas

Motivations and methods for text simplification

Long and complicated sentences prove to be a stumbling block for current systems relying on NL input. These systems stand to gain from methods that syntactically simplify such sentences. To simplify a sentence, we need an idea of the structure of the sentence, to identify the components to be separated out. Obviously a parser could be used to obtain the complete structure of the sentence. However, full parsing is slow and prone to failure, especially on complex sentences. In this paper, we consider two alternatives to full parsing which could be used for simplification. The first approach uses a Finite State Grammar (FSG) to produce noun and verb groups while the second uses a Supertagging model to produce dependency linkages. We discuss the impact of these two input representations on the simplification process.

Disambiguation of super parts of speech (or supertags): almost parsing

In a lexicalized grammar formalism such as Lexicalized Tree-Adjoining Grammar (LTAG), each lexical item is associated with at least one elementary structure (supertag) that localizes syntactic and semantic dependencies. Thus a parser for a lexicalized grammar must search a large set of supertags to choose the right ones to combine for the parse of the sentence. We present techniques for disambiguating supertags using local information such as lexical preference and local lexical dependencies. The similarity between LTAG and Dependency grammars is exploited in the dependency model of supertag disambiguation. The performance results for various models of supertag disambiguation such as unigram, trigram and dependency-based models are presented.

Automatic induction of rules for text simplification

Long and complicated sentences pose various problems to many state-of-the-art natural language technologies. We have been exploring methods to automatically transform such sentences in order to make them simpler. These methods involve the use of a rule-based system, driven by the syntax of the text in the domain of interest. Hand-crafting rules for every domain is time-consuming and impractical. The paper describes an algorithm and an implementation by which generalized rules for simplification are automatically induced from annotated training material using a novel partial parsing technique which combines constituent structure and dependency information. The algorithm described in the paper employs example-based generalizations on linguistically motivated structures.

XTAG system: a wide coverage grammar for English

This paper present the XTAG system, a grammar development tool based on the Tree Adjoining Grammar (TAG) formalism that includes a wide-coverage syntactic grammar for English. The various components of the system are discussed and preliminary evaluation results from the parsing of various corpora are given. Results from the comparison of XTAG against the IBM statistical parser and the Alvey Natural Language Tool parser are also given.

Almost parsing technique for language modeling

We present an approach that incorporates structural information into language models without really parsing the utterance. This approach brings together the advantages of an n-gram language model-speed, robustness and the ability to integrate with the speech recognizer with the need to model syntactic constraints under a uniform representation. We also show that our approach produces better language models than language models based on part-of-speech tags.

Glean: using syntactic information in document filtering

In this paper, we describe a system called Glean, which is based on the idea that coherent text contains signi cant latent information, such as syntactic structure and patterns of language use, which can be used to enhance the performance of information retrieval systems. We propose an approach to increase the precision of information retrieval that makes use of syntactic information obtained using a supertagger. In this approach, patterns based on local syntactic context are induced from training material. These patterns are used to re ne the set of documents retrieved by a standard Web search engine or an information retrieval system, by selecting relevant information and ltering out irrelevant items. We show that syntactic information does improve the e ectiveness of ltering irrelevant documents, and that supertagging is more e ective than part of speech tagging in ltering documents. Further, we also show how the extent of syntactic context a ects ltering performance. We discuss the relationship between Glean and other attempts at improving information retrieval performance.In this paper, we describe a system called Glean, which is based on the idea that coherent text contains significant latent information, such as syntactic structure and patterns of language use, which can be used to enhance the performance of information retrieval systems. We propose an approach to increase the precision of information retrieval that makes use of syntactic information obtained using a supertagger. In this approach, patterns based on local syntactic context are induced from training material. These patterns are used to refine the set of documents retrieved by a standard Web search engine or an information retrieval system, by selecting relevant information and filtering out irrelevant items. We show that syntactic information does improve the effectiveness of filtering irrelevant documents, and that supertagging is more effective than part of speech tagging in filtering documents. Further, we also show how the extent of syntactic context affects filtering performance. We discuss the relationship between Glean and other attempts at improving information retrieval performance.

EAGLE: An Extensible Architecture for General Linguistic Engineering

Over the course of two summer projects, we developed a general purpose natural language system which advances the state-of-the-art in several areas. The system contains demonstrated advancements in part-of-speech tagging, end-of-sentence detection, and coreference resolution. In addition, we believe that we have strong maximal noun phrase detection, and subject-verb-object recognition and a pat tern matching language well suited to a range of tasks. Other features of the system include modularity and interchangeability of components, rapid component integration and a debugging environment.

Explanation-based learning and finite state transducers: applications to parsing lexicalized tree adjoining grammars

There are currently two philosophies for building grammars and parsers: hand-crafted, wide coverage grammars; and statistically induced grammars and parsers. Aside from the methodological differences in grammar construction, the linguistic knowledge which is overt in the rules of handcrafted grammars is hidden in the statistics derived by probabilistic methods, which means that generalizations are also hidden and the full training process must be repeated for each domain. Although handcrafted wide coverage grammars are portable, they can be made more efficient when applied to limited domains, if it is recognized that language in limited domains is usually well constrained and certain linguistic constructions are more frequent than others. We view a domain-independent grammar as a repository of portable grammatical structures whose combinations are to be specialized for a given domain. We use Explanation-Based Learning (EBL) to identify the relevant subset of a handcrafted general purpose grammar (XTAG) needed to parse in a given domain (ATIS). We exploit the key properties of Lexicalized Tree-Adjoining Grammars to view parsing in a limited domain as finite state transduction from strings to their dependency structures.