A Best-First Search Algorithm for Generating Referring Expressions

Existing algorithms for generating referential descriptions to sets of objects have serious deficits: while incremental approaches may produce ambiguous and redundant expressions, exhaustive searches are computationally expensive. Mediating between these extreme control regimes, we propose a best-first searching algorithm for uniquely identifying sets of objects. We incorporate linguistically motivated preferences and several techniques to cut down the search space. Preliminary results show the effectiveness of the new algorithm.