... 157–164,Ann Arbor, June 2005.c2005 Association for Computational LinguisticsMachine Learning for Coreference Resolution: From Local Classification to Global RankingVincent NgHuman Language Technology ... need to make tough or potentiallysuboptimal design decisions.1 For instance, if we1We still need to determine the coreference systems to beemployed in our framework, however. Fortunately, ... fa-vorably to two state-of-the-art coreference systemsadopting the standard machine learning approach,outperforming them by as much as 4–7% on thethree data sets for one of the performance metrics.2...