The Annotation Graph Toolkit: Software Components for Building Linguistic Annotation Tools

The Annotation Graph Toolkit: Software Components for Building Linguistic Annotation Tools Kazuaki Maeda, Steven Bird, Xiaoyi Ma and Haejoong Lee Linguistic Data Consortium, University of Pennsylvania 3615 Market St., Philadelphia, PA 19104-2608 USA fmaeda, sb, xma, haejoongg@ldc.upenn.edu ABSTRACT Annotation graphs provide an efficient and expressive data model for linguistic annotations of time-series data. This paper reports progress on a complete software infrastructure supporting the rapid development of tools for transcribing and annotating time-series data. This general-purpose infrastructure uses annotation graphs as the underlying model, and allows developers to quickly create special-purpose annotation tools using common components. An application programming interface, an I/O library, and graphical user interfaces are described. Our experience has shown us that it is a straightforward task to create new special-purpose annotation tools based on this general-purpose infrastructure. Keywords transcription, coding, annotation graph, interlinear text, dialogue annotation 1. INTRODUCTION Annotation graphs (AGs) provide an efficient and expressive data model for linguistic annotations of time-series data [2]. This paper reports progress on a complete software infrastructure sup- porting the rapid development of tools for transcribing and anno- tating time-series data. This general-purpose infrastructure uses annotation graphs as the underlying model, and allows developers to quickly create special-purpose annotation tools using common components. This work is being done in cooperation with the developers of other widely used annotation systems, Transcriber and Emu [1, 3]. The infrastructure is being used in the development of a series of annotation tools at the Linguistic Data Consortium. Several such tools are shown in the paper: one for dialogue annotation, one for telephone conversation transcription, and one for interlinear transcription aligned to speech. This paper will cover the following points: the application pro- gramming interfaces for manipulating annotation graph data and importing data from other formats; the model of inter-component . communication which permits easy reuse of software components; and the design of the graphical user interfaces, which have been tailored to be maximally ergonomic for the tasks. The project homepage is: [http://www.ldc.upenn.edu/ AG/]. The software tools and software components described in this paper are available through a CVS repository linked from this homepage. 2. ARCHITECTURE 2.1 General Architecture Existing annotation tools are based on a two level model (Fig- ure 1 Top). The systems we demonstrate are based around a three level model, in which annotation graphs provide a logical level independent of application and physical levels (Figure 1 Bottom). The application level represents special-purpose tools built on top of the general-purpose infrastructure at the logical level. The system is built from several components which instantiate this model. Figure 2 shows the architecture of the tools currently being developed. Annotation tools, such as the ones discussed below, must provide graphical user interface components for signal visualization and annotation. The communication between compo- nents is handled through an extensible event language. An appli- cation programming interface for annotation graphs (AG-API) has been developed to support well-formed operations on annotation graphs. This permits applications to abstract away from file format issues, and deal with annotations …

Content truncated to 500 words