Randall A. Helzerman

Extensions to constraint dependency parsing for spoken language processing

A text-based and spoken language processing framework based on the constraint dependency grammar (CDG) developed by Maruyama is discussed. The scope of CDG is expanded to allow for the analysis of sentences containing lexically ambiguous words, to allow feature analysis in constraints, and to efficiently process multiple sentence candidates that are likely to arise in spoken language processing. The benefits of the CDG parsing approach are summarized. Additionally, the development of CDG grammars using our grammar tools and parser is discussed.

A parallel implementation of a hidden Markov model with duration modeling for speech recognition

This paper describes a parallel implementation of a Hidden Markov Model (HMM) for spoken language recognition on the MasPar MP-1. By exploiting the massive parallelism of explicit duration HMMs, we can develop more complex models for real-time speech recognition. Implementational issues such as choice of data structures, method of communication, and utilization of parallel functions are explored. The results of our experiments show that the parallelism in HMMs can be effectively exploited by the MP-1. Training that use to take nearly a week can now be completed in about an hour. The system can recognize the phones of a test utterance in a fraction of a second.<<ETX>>

Semantics and constraint parsing of word graphs

A constraint-based parser capable of processing a word graph containing multiple sentence hypotheses has been developed. When syntactic constraints are applied to a word graph, this parse is able to prune the graph of many ungrammatical sentence hypotheses and limit the possible parses of the remaining sentences. However, in many cases syntactic information alone is insufficient for selecting a single sentence hypothesis from a word graph. Hence, semantic constraints have been added to the parser to limit ambiguity further. The authors review the constraint parsing algorithm and then provide a simple example illustrating how syntactic and semantic features can be used to prune word candidates from a word graph and eliminate incorrect parses for the remaining sentences. They also report on the effectiveness of syntactic and semantic constraints for reducing the ambiguity of word networks constructed for N-best sentence hypotheses provided by the ATIS (Air Travel Information System) database.<<ETX>>

Implementation issues in the development of the PARSEC parser

This paper describes the implementation of a constraint‐based parser, PARSEC (Parallel ARchitecture SEntence Constrainer), which has the required flexibility that a user may easily construct a custom grammar and test it. Once the user designs grammar parameters, constraints, and a lexicon, our system checks them for consistency and creates a parser for the grammar. The parser has an X‐windows interface that allows a user to view the state of a parse of a sentence, test new constraints, and dump the constraint network to a file. The parser has an option to perform the computationally expensive constraint propagation steps on the MasPar MP‐1. Stream and socket communication was used to interface the MasPar constraint parser with a standard X‐windows interface on our Sun Sparcstation.

MUSE CSP: an extension to the constraint satisfaction problem

This paper describes an extension to the constraint satisfaction problem (CSP) called MUSE CSP (MUltiply SEgmented Constraint Satisfaction Problem). This extension is especially useful for those problems which segment into multiple sets of partially shared variables. Such problems arise naturally in signal processing applications including computer vision, speech processing, and handwriting recognition. For these applications, it is often difficult to segment the data in only one way given the low-level information utilized by the segmentation algorithms. MUSE CSP can be used to compactly represent several similar instances of the constraint satisfaction problem. If multiple instances of a CSP have some common variables which have the same domains and constraints, then they can be combined into a single instance of a MUSE CSP, reducing the work required to apply the constraints. We introduce the concepts of MUSE node consistency, MUSE arc consistency, and MUSE path consistency. We then demonstrate how MUSE CSP can be used to compactly represent lexically ambiguous sentences and the multiple sentence hypotheses that are often generated by speech recognition algorithms so that grammar constraints can be used to provide parses for all syntactically correct sentences. Algorithms for MUSE arc and path consistency are provided. Finally, we discuss how to create a MUSE CSP from a set of CSPs which are labeled to indicate when the same variable is shared by more than a single CSP.

Managing Multiple Knowledge Sources In Constraint-Based Parsing Of Spoken Language

In this paper, we describe a system which is capable of utilizing a variety of knowledge sources to select the most appropriate parse for a spoken sentence. These knowledge sources include syntax, semantics, and contextual information. We discuss one way to utilize contextual information when determining the parse for a sentence. At its simplest level, the system can be thought of as a general-purpose query answering system for multiple topical databases. The users input would be processed by the language processor which interfaces to the databases with the goal of interacting with the correct database in order to provide a reasonable answer to the users spoken request. Initially, it analyzes a word graph of sentence hypotheses provided by a speech recognizer using general syntactic and semantic rules. Then, if the utterance is still ambiguous, it utilizes context-specific constraints to further refine the analysis. This brings us closer to developing a more general purpose interface for multiple databases.

Parallel parsing of spoken language

The authors extended H. Maruyamas (1990) constraint dependency grammar (CDG) to process a lattice of sentence hypotheses instead of separate test strings. A postprocessor to a speech recognizer producing N-best hypotheses generates the word lattice representation, which is then augmented with information required for parsing. The authors summarize the CDG parsing algorithm and describe how the algorithm is extended to process the lattice on a single processor machine. They outline the CRCW P-RAM algorithm for parsing the word lattice, which requires O(n/sup 4/) processors to parse in O(k+n) time.<<ETX>>