Retriever: a self-training agent for intelligent informationdiscovery

摘要

With the exponential growth of the Internet and the volume ofinformation published on it, searching for information of interest hasbecome a very difficult and time-consuming task. In this paper, wepresent `Retriever', an autonomous agent that executes user queries andreturns high-quality results to the user. Retriever utilizes existingsearch engines to obtain the starting points for its subsequentautonomous exploration of the Web. Then it conducts a self-trainingprocess in order to learn the query domain and to increase itsefficiency. When the query domain is learned, the agent expands theoriginal query, reforms its search strategy and goes out looking for thedocuments to be presented to the user. It also incorporates relevancefeedback in order to perform subsequent searches on the same query