Composing and Organizing Language Services

Web service has gained its popularity in the past decade and emerged as a great impact on both industrial and research. The visionary promise of delivering dynamic creation of loosely coupled information system is almost into reality. This vision is called services computing. One of the wide implementations of Web services is language service. The number of language service available on the Web is inevitably increasing. Computer scientists collaborated with computational linguistics have been trying to develop more and more language services. When a new idea is proposed in services computing, it usually provides not enough detail of a domain used as a business case. It inspires us to look into particular service domain that can be later abstracted to other service domains. In this thesis, we describe our contributions to language services. We enhance the two basic parts in services computing, service composition and organization; and apply these to language services. There are some other elements in services computing, such as service monitoring and interfacing, but we believe that these two elements are the most crucial ones for language services. In service discovery, a preliminary process of service composition, we aim at providing a quantitative means of specifying the extent to which a provided service meets the requirements of a user’s request. This is important because it allows us to return a finer level of granularity in discovering language services. With the boom of language services, semantic similarity plays important role in discriminating services based on the language service ontology.