Charting Democracy Across Parsers
Scott Nowson, Robert Dale · 2007
Different parsers trained on the same corpus deliver different results, both in terms of overall performance and in terms of the individual analyses they provide. In particular, for any given sentence, one parser may provide a correct analysis, while another will produce an incorrect analysis; but when faced with a different sentence, the first parser may be in error while the second is correct. In this paper, we leverage this observation by exploring how the results of a number of different parsers may be combined to provide a better performance than any single parser. The method involves constructing a chart that contains edges contributed by a collection of parsers, with a simple voting mechanism to choose the most preferred constituents; this provides a significant improvement in performance over any individual parser. More sophisticated voting mechanisms are also discussed. 1