A Non-Monotonic Arc-Eager Transition System for Dependency Parsing
Matthew Honnibal, Yoav Goldberg, Mark S. Johnson · 2013
Previous incremental parsers have used monotonic state transitions. However, transitions can be made to revise previous decisions quite naturally, based on further information. We show that a simple adjustment to the Arc-Eager transition system to relax its monotonicity constraints can improve accuracy, so long as the training data includes examples of mistakes for the nonmonotonic transitions to repair. We evaluate the change in the context of a stateof-the-art system, and obtain a statistically significant improvement (p < 0.001) on the English evaluation and 5/10 of the CoNLL languages. 1