Joint learning improves semantic role labeling

Toutanova, Kristina; Haghighi, Aria; Manning, Christopher D.

doi:10.3115/1219840.1219913

Cited by 98 publications

(105 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Thus it is often beneficial to develop joint models to identify the various elements of a frame (Toutanova et al, 2005). However, these assumptions are less viable when dealing with emotions in tweets.…”

Section: Challenges Of Semantic Role Labeling Of Emotions In Tweetsmentioning

confidence: 99%

Semantic Role Labeling of Emotions in Tweets

Mohammad¹,

Zhu²,

Martin³

2014

Proceedings of the 5th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis

View full text Add to dashboard Cite

Past work on emotion processing has focused solely on detecting emotions, and ignored questions such as 'who is feeling the emotion (the experiencer)?' and 'towards whom is the emotion directed (the stimulus)?'. We automatically compile a large dataset of tweets pertaining to the 2012 US presidential elections, and annotate it not only for emotion but also for the experiencer and the stimulus. We then develop a classifier for detecting emotion that obtains an accuracy of 56.84 on an eight-way classification task. Finally, we show how the stimulus identification task can also be framed as a classification task, obtaining an F-score of 58.30.

show abstract

Section: Challenges Of Semantic Role Labeling Of Emotions In Tweetsmentioning

confidence: 99%

Semantic Role Labeling of Emotions in Tweets

Mohammad¹,

Zhu²,

Martin³

2014

Proceedings of the 5th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis

View full text Add to dashboard Cite

show abstract

“…One way to train these global features is to learn a linear classifier that selects a parse / frame pair from the ranked list, in the manner of Collins (2000). Reranking has previously been applied to semantic role labeling by Toutanova et al (2005), from which we use several features. The difference between this paper and Toutanova et al is that instead of reranking k-best SRL frames of a single parse tree, we are reranking 1-best SRL frames from the k-best parse trees.…”

Section: Training a Reranker Using Global Featuresmentioning

confidence: 99%

Joint parsing and semantic role labeling

Sutton

McCallum

2005

Proceedings of the Ninth Conference on Computational Natural Language Learning - CONLL '05

View full text Add to dashboard Cite

A striking feature of human syntactic processing is that it is context-dependent, that is, it seems to take into account semantic information from the discourse context and world knowledge. In this paper, we attempt to use this insight to bridge the gap between SRL results from gold parses and from automatically-generated parses. To do this, we jointly perform parsing and semantic role labeling, using a probabilistic SRL system to rerank the results of a probabilistic parser. Our current results are negative, because a locallytrained SRL model can return inaccurate probability estimates.

show abstract

“…This concept is intuitive when reasoning about the link between syntax and semantics, and it has been used earlier in semantic interpreters such as Absity (Hirst, 1983). However, except from a few tentative experiments (Toutanova et al, 2005), grammatical function is not explicitly used by current automatic SRL systems, but instead emulated from constituent trees by features like the constituent position and the governing category. More generally, these linguistic reasons have made a number of linguists argue that dependency structures are more suitable for explaining the syntax-semantics interface (Mel'čuk, 1988;Hudson, 1984).…”

Section: Introductionmentioning

confidence: 99%

The effect of syntactic representation on semantic role labeling

Johansson

Nugues

2008

Proceedings of the 22nd International Conference on Computational Linguistics - COLING '08

View full text Add to dashboard Cite

Almost all automatic semantic role labeling (SRL) systems rely on a preliminary parsing step that derives a syntactic structure from the sentence being analyzed. This makes the choice of syntactic representation an essential design decision. In this paper, we study the influence of syntactic representation on the performance of SRL systems. Specifically, we compare constituent-based and dependencybased representations for SRL of English in the FrameNet paradigm.Contrary to previous claims, our results demonstrate that the systems based on dependencies perform roughly as well as those based on constituents: For the argument classification task, dependencybased systems perform slightly higher on average, while the opposite holds for the argument identification task. This is remarkable because dependency parsers are still in their infancy while constituent parsing is more mature. Furthermore, the results show that dependency-based semantic role classifiers rely less on lexicalized features, which makes them more robust to domain changes and makes them learn more efficiently with respect to the amount of training data.

show abstract

Joint learning improves semantic role labeling

Cited by 98 publications

References 13 publications

Semantic Role Labeling of Emotions in Tweets

Semantic Role Labeling of Emotions in Tweets

Joint parsing and semantic role labeling

The effect of syntactic representation on semantic role labeling

Contact Info

Product

Resources

About