Machine Translation 16th China Conference, CCMT 2020, Hohhot, China, October 10-12, 2020, Revised Selected Papers

Preface\nOrganization\nContents\nTransfer Learning for Chinese-Lao Neural Machine Translation with Linguistic Similarity\n	Abstract\n	1 Introduction\n	2 Linguistic Similarity Between Thai and Lao\n	3 Our Approach\n		3.1 Chinese-Thai NMT Model\n		3.2 Thai-Lao NMT Model\n		3.3 Chinese-Lao NMT Model\n	4 Evaluation\n		4.1 Experimental Setup\n		4.2 Experimental Results\n	5 Related Work\n	6 Conclusions\n	Acknowledgements\n	References\nMTNER: A Corpus for Mongolian Tourism Named Entity Recognition\n	Abstract\n	1 Introduction\n	2 Related Work\n	3 Challenge for Mongolian Tourism NER\n	4 Annotated Mongolian Tourism Corpus\n		4.1 Data Collection\n		4.2 Annotation Schema\n		4.3 Annotation Agreement\n	5 Mongolian Tourism NER Model\n	6 Experiment\n		6.1 Data\n		6.2 Baselines\n		6.3 Results\n		6.4 Analysis\n	7 Conclusion\n	Acknowledgement\n	References\nUnsupervised Machine Translation Quality Estimation in Black-Box Setting\n	Abstract\n	1 Introduction\n	2 Background\n		2.1 Machine Translation Evaluation\n		2.2 Machine Translation Quality Estimation\n	3 Model Description\n		3.1 Pretrained Models for Quality Estimation\n		3.2 MTE-Based QE Data\n	4 Experiment\n		4.1 Setup\n		4.2 Experiment Results\n	5 Analysis\n		5.1 Is BERT Always the Best?\n		5.2 Is Black-Box Model Necessary?\n		5.3 Where Is the Limitation of QE?\n	6 Conclusion\n	Acknowledgement\n	References\nYuQ: A Chinese-Uyghur Medical-Domain Neural Machine Translation Dataset Towards Knowledge-Driven\n	1 Introduction\n	2 Related Work\n	3 Datasets\n		3.1 Data Collection\n		3.2 Corpus Preprocessing\n		3.3 Annotation\n		3.4 Knowledge Graph Construction\n	4 Corpus Analysis\n		4.1 Lexical Feature Analysis\n		4.2 Contrastive Analysis of Lexical Features\n	5 Experiments\n		5.1 Models\n		5.2 Setup\n		5.3 Automatic Evaluation\n		5.4 Manual Evaluation\n		5.5 Metrics\n		5.6 Annotation Statistics\n		5.7 Results\n		5.8 Case Study\n		5.9 Ablation Study\n	6 Conclusion and Future Work\n	References\nQuality Estimation for Machine Translation with Multi-granularity Interaction\n	1 Introduction\n	2 Related Work\n	3 Methodology\n		3.1 Model Architecture\n		3.2 Multi-granularity Interaction\n		3.3 Model Training\n	4 Experiments\n		4.1 Dataset\n		4.2 Experimental Setup\n		4.3 Experimental Result\n		4.4 Word-Level Feature Analysis\n	5 Conclusion\n	References\nTransformer-Based Unified Neural Network for Quality Estimation and Transformer-Based Re-decoding Model for Machine Translation\n	1 Introduction\n	2 Model\n		2.1 Transformer-Based Unified Neural Network for the Quality Estimation of Machine Translation\n		2.2 Study of Re-decoding-Based Neural Machine Translation\n	3 Experiment\n		3.1 Setting\n		3.2 Results\n		3.3 Analysis\n	4 Conclusions\n	References\nNJUNLP\'s Machine Translation System for CCMT-2020 Uighur  Chinese Translation Task\n	1 Introduction\n	2 Machine Translation System\n		2.1 Pre-processing\n		2.2 Architecture\n		2.3 Back-Translation of Monolingual Data\n		2.4 Fine-Tuning\n		2.5 Ensemble Translation\n		2.6 Reranking\n	3 Results\n	4 Conclusion\n	References\nDescription and Findings of OPPO\'s Machine Translation Systems for CCMT 2020\n	1 Introduction\n	2 Applying Multiple Word Segmentation Tools\n	3 English  Chinese Machine Translation Task\n		3.1 Data Preprocessing\n		3.2 Model Training\n		3.3 Corpus Filtering Task\n	4 Japanese  English Translation Task (Patent Domain)\n		4.1 Data Preprocessing\n		4.2 Model Training\n	5 Minority Languages  Mandarin Translation Task\n		5.1 Data Preprocessing\n		5.2 Model Training\n	6 Conclusion and Future Work\n	References\nTsinghua University Neural Machine Translation Systems for CCMT 2020\n	1 Introduction\n	2 Methods\n		2.1 Data\n		2.2 Models\n		2.3 Data Augmentation\n		2.4 Finetuning\n		2.5 Ensemble\n	3 Experiments\n		3.1 Settings\n		3.2 Results on Chinese-English Translation\n		3.3 Results on English-Chinese Translation\n	4 Conclusion\n	References\nBJTU’s Submission to CCMT 2020 Quality Estimation Task\n	Abstract\n	1 Introduction\n	2 Model Description\n		2.1 Pretrained Models for Quality Estimation\n		2.2 Further Pretraining for Bilingual Input\n		2.3 Multi-task Learning for Multi-granularities\n		2.4 Weighted Loss for Unbalanced Word Labels\n		2.5 Multi-model Ensemble\n	3 Experiment\n		3.1 Dataset\n		3.2 Experiment Results\n		3.3 Ablation Study\n	4 Conclusion\n	Acknowledgement\n	References\nNJUNLP\'s Submission for CCMT20 Quality Estimation Task\n	1 Introduction\n	2 Methods\n		2.1 Existing Methods\n		2.2 Proposed Methods\n	3 Experiments\n		3.1 Dataset\n		3.2 Settings\n		3.3 Single Model Results\n		3.4 Ensemble\n	4 Analysis\n	5 Conclusion\n	References\nTencent Submissions for the CCMT 2020 Quality Estimation Task\n	1 Introduction\n	2 Architecture\n		2.1 Predictors\n		2.2 Estimators\n		2.3 Ensemble\n	3 Experiments and Results\n		3.1 Dataset\n		3.2 Experiments\n	4 Conclusion\n	References\nNeural Machine Translation Based on Back-Translation for Multilingual Translation Evaluation Task\n	Abstract\n	1 Introduction\n	2 Related Work\n	3 Model\n		3.1 Transformer-Base\n		3.2 Transformer-Big\n		3.3 Dynamic-Conv\n	4 Experiments\n		4.1 Preprocessing\n		4.2 Back Translation Based Synthetic Data\n		4.3 Multi-model Ensemble\n		4.4 Contrast Experiments\n		4.5 Results\n		4.6 Model Analysis and Discussion\n	5 Conclusion and Future Work\n	Acknowledgement\n	References\nAuthor Index