A Unifying Theory of Transition-based and Sequence Labeling Parsing

2020/11/01 by Gómez-Rodríguez, Carlos, Strzyz, Michalina, Vilares, David
#68Q45 #68T50 #Computation and Language (cs.CL) #F.4.3 #FOS: Computer and information sciences #Formal Languages and Automata Theory (cs.FL) #I.2.7

paper · doi:10.48550/arxiv.2011.00584

Abstract

We define a mapping from transition-based parsing algorithms that read sentences from left to right to sequence labeling encodings of syntactic trees. This not only establishes a theoretical relation between transition-based parsing and sequence-labeling parsing, but also provides a method to obtain new encodings for fast and simple sequence labeling parsing from the many existing transition-based parsers for different formalisms. Applying it to dependency parsing, we implement sequence labeling versions of four algorithms, showing that they are learnable and obtain comparable performance to existing encodings.

Related

• Viable Dependency Parsing as Sequence Labeling
• Not All Linearizations Are Equally Data-Hungry in Sequence Labeling Parsing
• Dependency Graph Parsing as Sequence Labeling
• Discontinuous Constituent Parsing as Sequence Labeling
• Left-to-Right Dependency Parsing with Pointer Networks