Controlled Document Authoring in a Machine Translation

Cover
Half Title
Series Page
Title Page
Copyright Page
Contents
Figures
Tables
Abbreviations
Preface
Acknowledgements
Part I Research background
	1 Introduction
		1.1 Background
		1.2 Problems
			1.2.1 Document-level issues
			1.2.2 Sentence-level issues
			1.2.3 Terminology issues
		1.3 Solution scenario
		1.4 Research questions
		1.5 Scope
		1.6 Chapter organisation
	2 Related work
		2.1 Document formalisation
			2.1.1 Existing framework for document formalisation
				2.1.1.1 DITA
				2.1.1.2 Other document frameworks
			2.1.2 Descriptive analysis of document
				2.1.2.1 Genre analysis
				2.1.2.2 Functional structure analysis
			2.1.3 Evaluation of document quality
		2.2 Machine translation
			2.2.1 Overview
				2.2.1.1 Architecture
				2.2.1.2 Difficulties
			2.2.2 Practical use of MT
				2.2.2.1 Source input control
				2.2.2.2 Customisation
				2.2.2.3 Target output control
			2.2.3 Evaluation of MT
				2.2.3.1 Human evaluation
				2.2.3.2 Automatic evaluation
		2.3 Controlled language
			2.3.1 CL for MT
			2.3.2 Japanese-based CL
			2.3.3 Document-level CL
			2.3.4 Parameters to be considered
			2.3.5 Formulation of CL
			2.3.6 Deployment of CL
			2.3.7 Evaluation of CL
		2.4 Terminology management
			2.4.1 Terminology and term
			2.4.2 Methods of terminology management
			2.4.3 Controlled terminology construction
				2.4.3.1 Corpus compilation
				2.4.3.2 Term extraction
				2.4.3.3 Term variation management
			2.4.4 Evaluation of terminology
		2.5 Support system for authoring and translation
			2.5.1 Document generation system
			2.5.2 Controlled authoring support
				2.5.2.1 Integrated environment
				2.5.2.2 CL authoring assistant
			2.5.3 Evaluation of authoring system
				2.5.3.1 Precision and recall
				2.5.3.2 Usability
Part II Controlled document authoring
	3 Document formalisation
		3.1 Analysis of functional document elements
			3.1.1 Documents to be analysed
			3.1.2 Functional elements of municipal procedures
		3.2 DITA specialisation
			3.2.1 Mapping of document elements
			3.2.2 Application of specialised DITA
		3.3 Summary
	4 Controlled language
		4.1 Formulation of CL rules
			4.1.1 Technical writing based CL rules
			4.1.2 Rewriting trial based CL rules
		4.2 Evaluation
			4.2.1 MT quality evaluation
				4.2.1.1 Experimental setup
				4.2.1.2 Overall result of MT quality
				4.2.1.3 Generally effective CL rules
				4.2.1.4 MT-dependent CL rules
				4.2.1.5 Optimal rules for each system
			4.2.2 Source-readability evaluation
				4.2.2.1 Experimental setup
				4.2.2.2 Results and discussions
			4.2.3 Compatibility of machine-translatability and source readability
		4.3 CL guidelines
			4.3.1 Selection of CL rules
			4.3.2 Detailed description of CL rules
		4.4 Summary
	5 CL contextualisation
		5.1 Context-dependent CL
		5.2 Resolving incompatibilities
		5.3 Preliminary investigation on pre-translation processing
			5.3.1 Setup
			5.3.2 Results and analysis
				5.3.2.1 Rule 1
				5.3.2.2 Rule 2
		5.4 Summary
	6 Terminology management
		6.1 Terminology construction
			6.1.1 Municipal parallel corpus
			6.1.2 Manual extraction of terms
				6.1.2.1 Terms to be captured
				6.1.2.2 Platform
				6.1.2.3 Term extraction and validation
				6.1.2.4 Results
			6.1.3 Typology of term variation
			6.1.4 Terminology control
		6.2 Coverage estimation
			6.2.1 Methods
				6.2.1.1 Self-referring coverage estimation
				6.2.1.2 Conditions for evaluation
				6.2.1.3 Expected number of terms
				6.2.1.4 Growth rate of lexical items
			6.2.2 Results and discussions
				6.2.2.1 Population types and present status of terminologies
				6.2.2.2 Terminology growth
		6.3 Summary
Part III MuTUAL: An authoringsupport system
	7 System development
		7.1 Design principle
		7.2 Modules
		7.3 Implementation
			7.3.1 Topic template
			7.3.2 CL authoring assistant
				7.3.2.1 Use scenario
				7.3.2.2 Implementation of CL rules and terminology
			7.3.3 MT and dictionary
			7.3.4 Automatic pre-/post-translation processing
			7.3.5 Similar text search
		7.4 Summary
	8 Evaluation of CL violation detection component
		8.1 Setup
		8.2 Overall results
		8.3 Detailed analysis
		8.4 Summary
	9 System usability evaluation
		9.1 Experimental setup
			9.1.1 Task design
				9.1.1.1 MT system and guideline
				9.1.1.2 CL rule implementation
				9.1.1.3 Data
				9.1.1.4 Condition
				9.1.1.5 Procedure
				9.1.1.6 Post-task questionnaire
				9.1.1.7 Follow-up interview
				9.1.1.8 MT evaluation
				9.1.1.9 ST evaluation
			9.1.2 Implementation
		9.2 Results and discussions
			9.2.1 Effectiveness
				9.2.1.1 Violation correction
				9.2.1.2 MT quality
				9.2.1.3 ST quality
			9.2.2 Efficiency
				9.2.2.1 Time
				9.2.2.2 Editing effort
			9.2.3 Satisfaction
				9.2.3.1 Satisfaction with the task
				9.2.3.2 Satisfaction with the system
			9.2.4 User feedback
			9.2.5 Detailed observation
				9.2.5.1 Variance across participants
				9.2.5.2 Transition of task time
		9.3 Summary
Part IV Conclusion
	10 Research findings and outlook
		10.1 Findings and achievements of the study
		10.2 Contributions and implications
			10.2.1 Document-level framework for extending MT and CL research
			10.2.2 From descriptive to prescriptive
		10.3 Remaining issues and future directions
	Appendices
Appendices
	A. CL guidelines (Japanese original version)
	B. CL guidelines (English translated version)
Bibliography
Index