Publications

Advanced search

Abstract

Daniel Ortiz-Martínez, Ismael García-Varea, Francisco Casacuberta. An empirical comparison of stack-decoding algorithms for statistical machine translation. Pattern Recognition and Image Analysis, First Iberian Conference, 2003. pp. 654-663. Springer-Verlag.

Unlike other heuristic search algorithms, stack-based decoders have been proved theoreticalyy to guarantee the avoidance of search errors in the decodinbg phase of a statistical machine translation system. The disadvantage of the stack-based decoders are the high computational requirements. Therefore, to make the decoding problem feasible for statistical machine translation, some heuristics optimations have to be performed. In this paper, we describe, study and implement the state of the art stack-based decoding algorithm for statistical machine translation, making an empirical comparison which focuses specifically on the optimazation problems, computational time and translation results. Results are also presented for two well known tasks, the TOURIST task and the HANSARDS task.