Information Retrieval - From Information Access to Contextual Retrieval

1 Introduction Information Retrieval (IR) deals with uncertainty and vagueness in information systems. Uncertainty is caused by the problem of representing the semantics of text and other media, which cannot be done in a perfect way. On the other hands, information needs to be answered by IR systems are often vague and cannot be specified precisely, thus leading to iterative query formulation. The generic IR task can be specified as " Retrieve that amount of knowledge which a user needs in a specific situation for solving his / her current problem " (Kuhlen 1991). This definition implies two major research issues: 1. IR should consider the specific user, the situation and the problem to be solved. This view leads to the notion of contextual retrieval. 2. For retrieving the necessary knowledge, all accessible knowledge sources should be exploited; which requires methods for global information access. These two topics also were the key issues describe in the final report of the workshop " Challenges in Information Retrieval and Language Modeling " (Allan & Croft 2003) which brought together 38 top-level researchers from the area of IR in fall 2002. In the remainder of this paper, we will first describe current research in the area of global information access (section 2), followed by the discussion of work on contextual retrieval (Section 3). Then we will discuss issues for further research, before coming to the final conclusions. Dieses Dokument wird unter folgender creative commons Lizenz veröffentlicht: