SteveCassidy

Title: SteveCassidy

Abstract: This research paper presents and harmonizes two independent efforts to model annotated speech databases, one at Macquarie University and one at the University of Pennsylvania. It discusses various query languages and applications, focusing on the annotation graph model. The research aims to develop platform-independent open-source tools for creating, browsing, searching, querying, and transforming linguistic databases, ultimately disseminating large linguistic databases over the internet.

Research Question: How can we develop a comprehensive and flexible model for annotated speech databases that can handle their multidimensional, heterogeneous, and dynamic nature, while also addressing the temporal complexity of the data?

Methodology: The paper proposes two database models: the Emu model from Macquarie University, which organizes data primarily in terms of its hierarchical structure, and the annotation graph model from the University of Pennsylvania, which foregrounds the temporal structure. The authors demonstrate the expressive equivalence of the two models.

Results: The research shows that both models can effectively represent annotated speech databases. The annotation graph model, however, is particularly well-suited for handling the temporal complexity of the data. It represents the data as a directed graph, with nodes representing annotations and edges representing temporal relationships between them.

Implications: The research has significant implications for the field of linguistics and natural language processing. It provides a robust and flexible framework for managing and querying annotated speech databases, which can be applied to a wide range of applications, such as automatic tagging and parsing, machine translation, and information retrieval. Furthermore, the open-source tools developed as part of this research can facilitate the sharing and dissemination of large linguistic databases across the internet.

In conclusion, the paper presents a comprehensive and flexible model for annotated speech databases that can effectively handle their multidimensional, heterogeneous, and dynamic nature, as well as their temporal complexity. The research has important implications for the field of linguistics and natural language processing, and the developed tools can facilitate the sharing and dissemination of large linguistic databases.

Link to Article: https://arxiv.org/abs/0204026v1 Authors: arXiv ID: 0204026v1