Variant Transduction: A Method for Rapid Development of Interactive Spoken Interfaces

We describe an approach (\variant transduction") aimed at reducing the e ort and skill involved in building spoken language interfaces. Applications are created by specifying a relatively small set of example utterance-action pairs grouped into contexts. No intermediate semantic representations are involved in the speci cation, and the con rmation requests used in the dialog are constructed automatically. These properties of variant transduction arise from combining techniques for paraphrase generation, classi cation, and examplematching. We describe how a spoken dialog system is constructed with this approach and also provide some experimental results on varying the number of examples used to build a particular application.