Risk Minimization and Language Modeling in Text Retrieval

With the dramatic increase in online information in recent years, text retrieval is becoming increasingly important. Although many different text retrieval approaches have been proposed and studied in the past decades, it is still a significant scientific challenge to develop principled retrieval approaches that also perform well empirically; so far, the theoretically well-motivated models have rarely led to good performance directly. It is also a great challenge in retrieval to develop models that may go beyond the traditional notion of topical relevance and capture more user factors, such as topical redundancy and sub-topic diversity.