Machine Translation: 17th China Conference, CCMT 2021, Xining, China, October 8–10, 2021, Revised Selected Papers

Preface
Organization
Contents
A Document-Level Machine Translation Quality Estimation Model Based on Centering Theory
	1 Introduction
	2 Related Work
	3 Centering Theory and Extraction of the Preferred Centers
		3.1 Centering Theory and Preferred Centers
		3.2 The Preferred Centers Extraction Model
		3.3 The Semi-supervised Preferred Center Extraction Method
	4 The Quality Estimation Model
		4.1 The Inner-Extractor
		4.2 The Outer-Extractor
		4.3 The Evaluator
	5 Experiments
		5.1 Metrics
		5.2 Dataset Description
		5.3 Preferred Centers Extraction
		5.4 QE Results
		5.5 Case Study
	6 Conclusion
	A Appendix
	B Appendix
	C Appendix
	References
SAU’S Submission for CCMT 2021 Quality Estimation Task
	1 Introduction
	2 Methods
		2.1 Basic Model
		2.2 Proposed Method
		2.3 Ensemble
	3 Experiment
		3.1 Dataset
		3.2 Settings
		3.3 Results of the Single Model
		3.4 Results of the Ensemble Methods
	4 Conclusion
	References
BJTU-Toshiba's Submission to CCMT 2021 QE and APE Task
	1 Introduction
	2 Chinese-English Sentence-Level Quality Estimation
		2.1 Model Description
		2.2 Multi-phase Pre-finetuning
		2.3 Partial-Input Estimation
		2.4 Model Ensemble
	3 Chinese-English Automatic Post-Editing
		3.1 BERT-initialized Transformer
		3.2 Domain Selection
		3.3 Data Augmentation Techniques
	4 Conclusion
	References
Low-Resource Neural Machine Translation Based on Improved Reptile Meta-learning Method
	1 Introduction
	2 Background
	3 Our Approach
		3.1 Unified Word Embedding Representation
		3.2 NMT Method Based on Improved Reptile Meta-learning
	4 Experiments
		4.1 Datasets
		4.2 Setting and Baseline
		4.3 Result and Analysis
		4.4 Ablation Experiments
		4.5 Case Study
	5 Conclusion
	References
Semantic Perception-Oriented Low-Resource Neural Machine Translation
	1 Introduction
	2 Background
	3 Method
		3.1 Semantic Perception-Assisted Pre-training Model
		3.2 Hierarchical Knowledge Distillation Training Process
	4 Experiments
		4.1 Datasets and Configuration
		4.2 Results and Analysis
		4.3 Ablation Experiment
		4.4 Case Study
	5 Conclusion
	References
Semantic-Aware Deep Neural Attention Network for Machine Translation Detection
	1 Introduction
	2 Related Work
	3 Model Overview
		3.1 Semantic-Aware Influencing Attention Network (SIAN) in Monolingual Scenario
		3.2 Semantic Consistency-Aware Interactive Attention Network (SCIA) in Bilingual Scenario
	4 Experiments
		4.1 Data Preparation
		4.2 Model Parameters Settings
		4.3 Evaluation Metric
		4.4 Model Comparison and Analysis in Monolingual Scenario
		4.5 Model Comparison and Analysis in the Bilingual Scenario
		4.6 Case Study
		4.7 Evaluation on Neural Machine Translation Systems
	5 Conclusion
	References
Routing Based Context Selection for Document-Level Neural Machine Translation
	1 Introduction
	2 Related Work
	3 Background
		3.1 Document-Level NMT
		3.2 Transformer
		3.3 Conditional Language-Specific Routing (CLSR)
	4 Method
		4.1 Inputs of Our Model
		4.2 Context Attention
		4.3 Auto-selection
	5 Experiments
		5.1 Datasets
		5.2 Training Detail
		5.3 Main Results
		5.4 Ablation Study
		5.5 Analysis
	6 Conclusion and Future Work
	References
Generating Diverse Back-Translations via Constraint Random Decoding
	1 Introduction
	2 Related Work
	3 Proposed Methods
		3.1 Fluency Boost Learning
		3.2 Evolution Decoding Algorithm
		3.3 Joint Training
	4 Experimental Setting
		4.1 Metrics
		4.2 Dataset
		4.3 Experiment Settings
	5 Results and Analysis
		5.1 Main Results
		5.2 Quantitative Analysis
		5.3 Qualitative Analysis
	6 Conclusion
	References
ISTIC's Neural Machine Translation System for CCMT' 2021
	1 Introduction
	2 System Architecture
		2.1 Baseline System
		2.2 Our System
	3 Methods in Different Tasks
		3.1 M2C Task, U2C Task, and T2C Task
		3.2 R2C Low Resource Task
	4 Experiments
		4.1 System Settings
		4.2 Data Pre-processing
		4.3 Experimental Results
	5 Conclusions
	References
BJTU's Submission to CCMT 2021 Translation Evaluation Task
	1 Introduction
	2 Data
		2.1 Chinese-English
		2.2 UyghurChinese
		2.3 TibetanChinese
	3 Model
	4 Method
		4.1 Data Augmentation
		4.2 Model Average
		4.3 Finetune
		4.4 Model Ensemble
		4.5 Reranking
	5 Experiment
		5.1 ChineseEnglish
		5.2 EnglishChinese
		5.3 UyghurChinese
		5.4 TibetanChinese
	6 Conclusion
	References
Author Index