The Necessity of Parsing for Predicate Argument Recognition

By Daniel Gildea, Martha Palmer. 2002

252 citations recorded by the source corpus.

Broad-coverage corpora annotated with semantic role, or argument structure, information are becoming available for the first time. Statistical systems have been trained to automatically label semantic roles from the output of statistical parsers on unannotated text. In this paper, we quantify the effect of parser accuracy on these systems' performance, and examine the question of whether a flatter "chunked" representation of the input can be as effective for the purposes of semantic role identification.

Subfields: Artificial Intelligence, Molecular Biology

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

In Molecular Biology

Citations in the corpus

What do you learn from context? Probing for sentence structure in contextualized word representations (2019, background): “Semantic role labeling beneﬁts greatly from contextual encoders overall, but this is predominantly due to better labeling of core roles (+19.0 F1 for ELMo) which are known to be closely tied to syntax (e.g. Punyakanok et al. (2008); Gildea & Palmer (2002)).”
The Proposition Bank: An Annotated Corpus of Semantic Roles (2005, positive): “Most recently, the Gildea and Palmer (2002) scores presented here have been improved markedly through the use of support-vector machines as well as additional features for named entity tags, headword POS tags, and verb clusters for back-off (Pradhan et al. 2003) and using maximum-entropy…”
Syntactic Scaffolds for Semantic Structures (2018, background): “…for a very tight integration of syntactic and semantic processing (Steedman, 2000; Copestake and Flickinger, 2000), and many systems have used syntactic dependency or phrase-based parsers as preprocessing for semantic analysis (Gildea and Palmer, 2002; Punyakanok et al., 2008; Das et al., 2014).”
A Classifier-Based Parser with Linear Run-Time Complexity (2005, background): “For example, the tree-path feature has been shown to be valuable in semantic role labeling (Gildea and Palmer, 2002).”
Dependency-based Semantic Role Labeling of PropBank (2008, background): “With very few exceptions (e.g. Collobert and Weston, 2007), published SRL methods have used some sort of syntactic structure as input (Gildea and Palmer, 2002; Punyakanok et al., 2008).”
NL-Augmenter: A Framework for Task-Sensitive Natural Language Augmentation (2021, unknown): “The transformation relies on propbank annotations (Bonial et al., 2012; Kingsbury and Palmer, 2002; Palmer et al., 2005; Gildea and Palmer, 2002).”
The Importance of Syntactic Parsing and Inference in Semantic Role Labeling (2008, methodology): “The work of Gildea and Palmer (2002) pioneered not only the fundamental architecture of SRL, but was also the ﬁrst to investigate the interesting question regarding the signiﬁcance of using full parsing for high quality SRL.”
Using Predicate-Argument Structures for Information Extraction (2003, methodology): “3 These results, listed also on the last line of of the decision-tree-based method against the re-sults obtained by the statistical approach reported in (Gildea and Palmer, 2002).”
The interaction of domain knowledge and linguistic structure in natural language processing: interpreting hypernymic propositions in biomedical text (2003, background): “Although these rules are generally eﬀective, the underspeciﬁed categorial analysis to which they apply does not provide detailed structural cues in support of argument identiﬁcation [55].”
The Effect of Syntactic Representation on Semantic Role Labeling (2008, background): “It has previously been shown that SRL systems need a syntactic structure as input (Gildea and Palmer, 2002; Punyakanok et al., 2008).”
A unified architecture for natural language processing: deep neural networks with multitask learning (2008, background): “This is an important result, given that the NLP community considers syntax as a mandatory feature for semantic extraction (Gildea & Palmer, 2001).”
Support Vector Learning for Semantic Argument Classification (2005, methodology): “Gildea & Palmer (2002) use the same features used by Gildea and Jurafsky (2002), which are also our initial set of features.”
Extracting Relations with Integrated Information Using Kernel Methods (2005, methodology): “Such analysis may be based on newly-annotated corpora like PropBank (Kingsbury and Palmer, 2002) at the University of Pennsylvania and NomBank (Meyers et al., 2004) at New York University.”
Intricacies of Collins’ Parsing Model (2004, result): “…(Gildea, 2001), sample selection (Hwa, 2001), bootstrapping non-English parsers (Hwa, Resnik, and Weinberg, 2002) and for the automatic labeling of semantic roles and predicate-argument extraction (Gildea and Jurafsky, 2000; Gildea and Palmer, 2002), as well as that of other research efforts.”
Calibrating Features for Semantic Role Labeling (2004, methodology): “The baseline system uses the original features proposed in (Gildea and Palmer, 2002) and each row shows the improvement over the baseline when that feature is added to the baseline features.”
Robust Parsing of the Proposition Bank (2006, background): “While some of these models are based on full parse trees (Gildea and Jurafsky, 2002; Gildea and Palmer, 2002), other methods have been proposed that eschew the need for a full 1 There are thirteen semantic role labels for modiﬁers.”