Computational Structures as Neural Symbolic Representation 公开

He, Han (Spring 2023)

Permanent URL: https://etd.library.emory.edu/concern/etds/r494vm50r?locale=zh
Published

Abstract

Although end-to-end neural models have been dominating Natural Language Processing for both performance and flexibility, critics have recently drawn attention to their poor generalization and lack of interpretability. Conversely, symbolic paradigms such as Abstract Meaning Representation (AMR) are humanly comprehensible but less flexible. In response, we propose Executable Abstract Meaning Representation (EAMR) as a reconciliation of both paradigms. EAMR is a neural symbolic framework that frames a task as a program, which interactively gets generated, revised and executed. In our novel definition, execution is a sequence of transforms on AMR graphs. Through a hybrid runtime, EAMR learns the automatic execution of AMR graphs, yet it also allows for the integration of handcrafted heuristics, knowledge bases and APIs. EAMR can be used in many applications such as dialogue understanding and response generation.

Table of Contents

1     Introduction                                                                                                     1

2     Background                                                                                                     4

2.1     Literature Review................................................................................. 4

2.2     Past Research........................................................................................ 6

3     Approach                                                                                                         9

3.1     Definition.............................................................................................. 9

3.2     Baselines............................................................................................. 10

3.3     Dialog Context Representation............................................................ 11

3.4     Step-by-Step Execution...................................................................... 13

3.5     Edit Oracle.......................................................................................... 14

3.6     AMR Oracle....................................................................................... 19

4     Experiments                                                                                                   21

4.1     Dataset..................................................................................................21

4.2     Metric................................................................................................... 22

4.3     Training............................................................................................... 23

4.4     Results.................................................................................................. 23

5     Application                                                                                                    26

5.1     Controllable Execution........................................................................ 26

5.2     Case Study........................................................................................... 28

6     Conclusion                                                                                                     31

Appendix A - Hyper-Parameter Configuration  33

About this Dissertation

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
Department
Degree
Submission
Language
  • English
Research Field
关键词
Committee Chair / Thesis Advisor
Committee Members
最新修改

Primary PDF

Thumbnail Title Date Uploaded Actions
Computational Structures as Neural Symbolic Representation () 2023-04-09 13:30:37 -0400

Supplemental Files

Thumbnail Title Date Uploaded Actions