Listwise Ranking Functions for Statistical Machine Translation

Meng Zhang,Yang Liu,Huanbo Luan,Maosong Sun

DOI: https://doi.org/10.1109/taslp.2016.2560527

2016-01-01

IEEE/ACM Transactions on Audio Speech and Language Processing

Abstract:Decision rules play an important role in the tuning and decoding steps of statistical machine translation. The traditional decision rule selects the candidate with the greatest potential from a candidate space by examining each candidate individually. However, viewing each candidate as independent imposes a serious limitation on the translation task. We instead view the problem from a ranking perspective that naturally allows the consideration of an entire list of candidates as a whole through the adoption of a listwise ranking function. Our shift from a pointwise to a list-wise perspective proves to be a simple yet powerful extension to current modeling that allows arbitrary pairwise functions to be incorporated as features, whose weights can be estimated jointly with traditional ones. We further demonstrate that our formulation encompasses the minimum Bayes risk (MBR) approach, another decision rule that considers restricted listwise information, as a special case. Experiments show that our approach consistently outperforms the baseline and MBR methods across the considered test sets.