Universal Dependencies (UD) is becom ing a standard annotation scheme cross-linguistically, but it is argued that this scheme centering on content words is harder to parse than the conventional one centering on function words. To improve the parsability of UD, we propose a back-and-forth conversion algorithm, in which we preprocess the training treebank to in crease parsability, and reconvert the parser outputs to follow the UD scheme as a post-process. We show that this technique con sistently improves LAS across languages even with a state-of-the-art parser, in par ticular on core dependency arcs such as nominal modifier. We also provide an in-depth analysis to understand why our method increases parsability.
展开▼
