When is text generation useful?
The TAILOR Approach
TAILOR is a text generation system which works within the domain of patents and it is designed to produce descriptions of objects when a query is posed to it by a user. The knowledge base it uses is called RESEARCHER and this is a program which is able to read, remember and generalize from patent abstracts. Obviously, the ability of TAILOR to generate text efficiently hinges on the success rate of RESEARCHER in executing its duties of parsing and generalizing. However, the study of RESEARCHER is subsumed under the school of knowledge representation and is therefore not the focus of this paper. It is assumed that RESEARCHER is adequate in its task execution and the analysis is focused primarily on the performance of TAILOR as a text generation system.
Generating descriptions is very difficult for TAILOR cause it has no parameters on how much information should be included in the answer. It is not a case of simple retrieval of information from RESEARCH. As mentioned above, there are criteria of lexis, discourse appropriation, functionality, content and presentation to fulfill. All these parameters are inferred from the User Model which is discussed in the following section.
As computer technology reaches ever more users in ever more domains, we anticipate many more interesting and practical applications for text generation. The advent of the World Wide Web, in particular, presents a wealth of opportunities, as more people can go to more places for more kinds of information than ever before. Text generation technology will allow the most up-to-date information to be communicated to the broadest audience, with the highest accuracy and quality of presentation.
Click here for a detailed look at the Tailor approach.
Conclusion