Reactive Control of Dynamic Progressive Processing

Abstract

Progressive processing is a model of computation that allows a system to tradeoff computational resources against the quality of results. This paper generalizes the existing model to make it suitable for dynamic composition of information retrieval techniques. The new framework addresses effectively the uncertainty associated with the duration and output quality of each component. We show how to construct an optimal meta-level controller for a single task based on solving a corresponding Markov decision problem, and how to extend the solution to the case of multiple and dynamic tasks using the notion of an opportunity cost. 1 Introduction This paper is concerned with the design and implementation of an effective computational model for information retrieval search engines that can tradeoff computational resources against the quality of the result. Our approach is based on run-time monitoring of the information retrieval process and on dynamic selection of information ret...