Speechbuilder: facilitating spoken dialogue system development

Abstract

SPEECHBUILDER is a suite of tools that helps facilitate the creation of mixed-initiative spoken dialogue systems for both novice and experienced developers of human language applications. SPEECHBUILDER employs intuitive methods of specification to allow developers to create human language interfaces to structured information stored in a relational database, or to control-and transaction-based applications. The goal of this project has been both to robustly accommodate the various scenarios where spoken dialogue systems may be needed, and to provide a stable and reliable infrastructure for design and deployment of applications. SpeechBuilder has been used in various spoken language domains, including a directory of the people working at the MIT Laboratory for Computer Science, an application to control the various physical items in a typical office environment, and a system for real-time weather information access.

A context resolution server for the galaxy conversational systems

“…First, it displays a significant amount of domain-dependent code and it is not very extensible. This component also does not support context resolution based on key-value pairs, which are the basis for knowledge representation in SPEECHBUILDER [19], a new software tool developed by SLS, which / 4…”

“…The intent of this thesis is to provide the reader with an understanding of the necessity for the new CR server and how it handles context resolution within our current domains. These include JUPITER [53] for weather information retrieval, MERCURY [44] for flight reservations, ORION [41] for online task delegation, PEGASUS [54] for flight arrival and departure information retrieval, and VOYAGER [17] for Boston area traffic and landmark information retrieval, as well as the simpler key-value-based domains created using the SPEECHBUILDER utility [19]. No meal is served on the first leg, and dinner is served on the second leg.…”

Automatic induction of n-gram language models from a natural language grammar

“…This framework can also be utilized to reduce the language model compilation time, which is almost entirely accounted for by the computation involved in parsing a large training corpus. This is especially important when delay becomes an issue, for example, in an interactive environment such as SPEECH-BUILDER [2]. An interesting solution is to parse only a small subset of the corpus to bootstrap the mapping transducer, and use FST operations outlined above to label the remaining sentences, i.e., treating them as if they failed to parse.…”

“…While this approach gives a performance that compares favorable to that of a standard class n-gram, it too suffers computationally, in part because our FST technology for speech recognition has been optimized for performance on traditional class n-gram's. An alternative approach, used by the SPEECHBUILDER [2] utility, is to generate both an NL grammar and a class n-gram from a set of annotated sentences. This paper details our work in developing a technique which satisfies all of the required constraints: computational efficiency, ease of maintenance, and high performance.…”

Development of a stochastic dialog manager driven by semantics

“…These systems are characterised by the following: telephone access, which is practical, tasks restricted to specific domain which limits the vocabulary, and mixed initiative. Descriptions of some of these systems which have been developed in the last years can be found in [1], [2], [3], [4], [5] y [6].…”