A Latent Variable Model of Synchronous Parsing for Syntactic and Semantic Dependencies

By James Henderson, Paola Merlo, Gabriele Antonio Musillo, Ivan Titov. 2008

61 citations recorded by the source corpus.

We propose a solution to the challenge of the CoNLL 2008 shared task that uses a generative history-based latent variable model to predict the most likely derivation of a synchronous dependency parser for both syntactic and semantic dependencies. The submitted model yields 79.1% macro-average F1 performance, for the joint task, 86.9% syntactic dependencies LAS and 71.0% semantic dependencies F1. A larger model trained after the deadline achieves 80.5% macro-average F1, 87.6% syntactic dependencies LAS, and 73.1% semantic dependencies F1.

Subfields: Artificial Intelligence

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Encoding Sentences with Graph Convolutional Networks for Semantic Role Labeling (2017, background): “Though historically most SRL approaches did rely on syntax (Thompson et al., 2003; Pradhan et al., 2005; Punyakanok et al., 2008; Johansson and Nugues, 2008), the last generation of SRL models put syntax aside in favor of neural sequence models, namely LSTMs (Zhou and Xu, 2015; Marcheggiani et al.,…”
A Latent Variable Model for Generative Dependency Parsing (2007, methodology): “Recent work has extended this ISBN model of dependency parsi ng to joint dependency parsing and semantic role labeling [19].”
Incremental Sigmoid Belief Networks for Grammar Learning (2010, methodology): “For example, ISBNs have been applied to the dependency parsing problem (Titov and Henderson, 2007) and to joint dependency parsing and semantic role labelling (Henderson et al., 2008; Gesmundo et al., 2009), where in both cases they achieved state-of-the-art results.”
The CoNLL 2008 Shared Task on Joint Parsing of Syntactic and Semantic Dependencies (2008, background): “The system of Henderson et al. (2008), which was designed for joint learning of the complete task, improves its rank from eighth to ﬁfth compared to the ofﬁcial results (Table 10).”
A Latent Variable Model of Synchronous Syntactic-Semantic Parsing for Multiple Languages (2009, methodology): “We participate in this task with a generative, history-based model proposed in the CoNLL 2008 0 Authors in alphabetical order. shared task for English (Henderson et al., 2008) and further improved to tackle non-planar dependencies (Titov et al., 2009).”
Greedy, Joint Syntactic-Semantic Parsing with Stack LSTMs (2016, methodology): “3 Constraints on Transitions To ensure that the parser never enters an invalid state, the sequence of transitions is constrained, following Henderson et al. (2013). Actions that copy or move items from the buffer (S-SHIFT, S-RIGHT and M-SHIFT) are forbidden when the buffer is empty.”
A Simple and Accurate Syntax-Agnostic Neural Model for Dependency-based Semantic Role Labeling (2017, methodology): “Similarly to Kiperwasser and Goldberg (2016) we used word dropout (Iyyer et al., 2015); we replace a word with the unknown token UNK with probability α fr ( w )+ α , where α is an hyper-parameter and fr ( w ) is the frequency of the word w .”
A Joint Model for Extended Semantic Role Labeling (2011, methodology): “Henderson et al. (2008) was designed for jointly learning to predict syntactic and semantic dependencies.”