We present a simple and effective framework for exploiting multiple monolingual treebanks with different annotation guidelines for parsing. Several types of transformation patterns (TP) are designed to capture the systematic annotation inconsistencies among different treebanks. Based on such TPs, we design quasisynchronous grammar features to augment the… (More)
Table 4: Parsing accuracy (UAS) comparison on CTB5test with gold-standard POS tags. Li11 refers to the second-order graph-based model of Li et al. (2011), whereas Z&N11 is the feature-rich transition-based model of Zhang and Nivre (2011).