A summary based language retrieval method

The performance of language retrieval method is largely determined by the accuracy of document language model. Motivated by the hypothesis that query-biased summary presents the information that is most relevant to a query, we propose a summary-biased approach to study the use of internal structures for the estimationod document language model. This method can be viewed as a two-stage language model, in the first stage, the document language model is smoothed by the query-biased summary, and in the second stage, the smoothed document language model is further smoothed by the collection language model. Moreover, our method can be used in feedback technique to estimate an improved query model.