When is Self-Training Effective for Parsing?

摘要

Self-training has been shown capable of improving on state-of-the-art parser performance (McClosky et al., 2006) despite the conventional wisdom on the matter and several studies to the contrary (Charniak, 1997; Steedman et al., 2003). However, it has remained unclear when and why self-training is helpful. In this paper, we test four hypotheses (namely, presence of a phase transition, impact of search errors, value of non-generative reranker features, and effects of unknown words). From these experiments, we gain a better understanding of why self-training works for parsing. Since improvements from self-training are correlated with unknown bi-grams and biheads but not unknown words, the benefit of self-training appears most influenced by seeing known words in new combinations.