Information-theoretical analysis of index searching: Revised

We present an information-thoeretical viewpoint for similarity-based retrieval along with index structures. This retrieval system has two stages: pruning data items based on the index structures, and matching surviving data items. The first stage is modeled as the Wyner-Ziv problem, while the second stage is considered as a coding problem where some of the decording results are available as partial side information at both the encoder and decoder. We clarify upper and lower bounds of the optimal retrieval performance. This also implies certain relationships between retrieval parameters and performance. Keywords— index searching, multiterminal source coding, rate-distortion theory, cascading, feedback