WISeN: Widely Interpretable Semantic Network for Richer Meaning Representation Public

Abstract

Many semantic annotations currently utilize Abstract Meaning Representation and PropBank frameset files to represent meaning. This scheme relies on arbitrary predicate-argument structures comprising unintuitive numbered arguments, fine-grained sense-disambiguation, and high start-up costs. To address these issues, we present a new annotation scheme, WISeN, that prioritizes semantic roles over numbered arguments and does away with sense-disambiguation. This scheme aims to be more intuitive for annotators and more interpretable by parsers. We evaluate this annotation scheme with a two-part experiment. First, we measure speed and accuracy of manual annotations. Second, we train a parser on both AMR and WISeN annotations and measure model accuracy. The results show that WISeN supports improved parser performance and increased inter-annotator agreement without sacrificing annotation speed compared to AMR. As such, we advocate for the adoption of WISeN as an annotation scheme for semantic representations.

Table of Contents

1 Introduction ................................................................ 1

2 Background, Related Work, & Rationale ......................... 4

2.1 Semantic Representations of Language ....................... 4

2.2 Thematic Roles ......................................................... 6

2.2.1 Semantics of Numbered Arguments ......................... 6

2.2.2 VerbNet Thematic Roles ......................................... 10

2.3 Sense Disambiguation .............................................. 14

2.4 Abstract Meaning Representation .............................. 15

2.5 AMR Parsing ............................................................ 17

2.6 Rationale ................................................................. 18

3 Methodology .............................................................. 23

3.1 Overview ................................................................. 23

3.2 Annotation Experiment ............................................ 25

3.2.1 Corpus .................................................................. 25

3.2.2 Annotators ........................................................... 26

3.2.3 Procedure ............................................................. 27

3.2.4 Evaluation Metrics ................................................ 28

3.3 Parsing Experiment ................................................. 28

3.3.1 AMR Corpus ......................................................... 29

3.3.2 WISeN Corpus ...................................................... 30

3.3.3 Parsing AMR and WISeN ....................................... 36

3.3.4 Evaluation Metrics ................................................ 39

4 Results ...................................................................... 41

4.1 Annotation Experiment .......................................... 41

4.1.1 Inter-Annotator Agreement ................................... 41

4.1.2 Speed of Annotations ........................................... 44

4.2 Parsing Experiment ................................................ 45

5 Discussion ................................................................ 49

5.1 WISeN Annotation is More Accurate than AMR ......... 49

5.2 WISeN Annotation is Comparable in Speed to AMR .. 50

5.3 WISeN Improves Parser Performance ....................... 51

5.4 Limitations ............................................................ 52

5.5 Future Work .......................................................... 55

5.6 Conclusions .......................................................... 56

About this Honors Thesis

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Emory College
Department	Mathematics and Computer Science
Degree	B.S.
Submission	Honors Thesis
Language	English
Research Field	Language, Linguistics Computer Science
Mot-clé	annotation parsing natural language processing
Committee Chair / Thesis Advisor	Choi, Jinho, Emory University
Committee Members	Kim, Yun, Emory University El-Sayed, Nosayba, Emory University

Dernière modification

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	WISeN: Widely Interpretable Semantic Network for Richer Meaning Representation ()	2021-04-12 23:08:31 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions