Implicit information retrieval

Abstract

In an information retrieval system, many queries depend upon the realization of information implicit in the information base for a satisfactory response. Such problems as the recognition of morphological variation in language, the detection of word or sentence synonymy in context and the determination of simple logical or causal consequences of the information base, all depend to a lesser or greater extent upon information not explicitly available by currently used techniques. We describe-an experimental model for the exploration of "implicit retrieval". Simple English text is mapped into a formalism suitable for machine inference. The performance of the model will be illustrated with extracts from the Wall Street Journal. Future application of the principles of implicit retrieval to conventional retrieval systems will be briefly discussed. © 1968.