Masao Utiyama

Zhuosheng Zhang, Kehai Chen, Rui Wang et al.

Representation learning is the foundation of natural language processing (NLP). This work presents new methods to employ visual information as assistant signals to general NLP tasks. For each sentence, we first retrieve a flexible number of images either from a light topic-image lookup table extracted over the existing sentence-image pairs or a shared cross-modal embedding space that is pre-trained on out-of-shelf text-image pairs. Then, the text and images are encoded by a Transformer encoder and convolutional neural network, respectively. The two sequences of representations are further fused by an attention layer for the interaction of the two modalities. In this study, the retrieval process is controllable and flexible. The universal visual representation overcomes the lack of large-scale bilingual sentence-image pairs. Our method can be easily applied to text-only tasks without manually annotated multimodal parallel corpora. We apply the proposed method to a wide range of natural language generation and understanding tasks, including neural machine translation, natural language inference, and semantic similarity. Experimental results show that our method is generally effective for different tasks and languages. Analysis indicates that the visual signals enrich textual representations of content words, provide fine-grained grounding information about the relationship between concepts and events, and potentially conduce to disambiguation.

Christoph Leiter, Haiyue Song, Hour Kaing et al.

Evaluating the cultural awareness of large language models is crucial to ensure the fairness of generated text and the generalizability of applications across the world. Recent benchmarks explore cultural goods like food or values like behavior in stressful situations through the lens of question answering or text generation tasks. However, creating these benchmarks requires time-intensive and costly human annotations. Also, benchmarks that evaluate cultural awareness in free text are scarce and often rely on dated evaluation mechanisms. To address this gap, we introduce ExCAM, an Explainable Cultural Awareness Metric, which is, to our knowledge, the first dedicated evaluation metric that identifies, rates and explains cultural errors in instruction-output pairs. To train and evaluate ExCAM, we introduce ExCAM40k, a dataset comprised of nine existing benchmarks that we reformat and enhance with synthetic errors. Compared to several baselines, including GPT-5, ExCAM achieves the highest error detection rate with up to 80% accuracy on a balanced test set. Therefore, ExCAM opens the pathway towards fine-grained and explainable cultural evaluation of free text.

Zhuosheng Zhang, Hai Zhao, Masao Utiyama et al.

Discriminative pre-trained language models (PLMs) learn to predict original texts from intentionally corrupted ones. Taking the former text as positive and the latter as negative samples, the PLM can be trained effectively for contextualized representation. However, the training of such a type of PLMs highly relies on the quality of the automatically constructed samples. Existing PLMs simply treat all corrupted texts as equal negative without any examination, which actually lets the resulting model inevitably suffer from the false negative issue where training is carried out on pseudo-negative data and leads to less efficiency and less robustness in the resulting PLMs. In this work, on the basis of defining the false negative issue in discriminative PLMs that has been ignored for a long time, we design enhanced pre-training methods to counteract false negative predictions and encourage pre-training language models on true negatives by correcting the harmful gradient updates subject to false negative predictions. Experimental results on GLUE and SQuAD benchmarks show that our counter-false-negative pre-training methods indeed bring about better performance together with stronger robustness.

Hour Kaing, Raj Dabre, Haiyue Song et al.

This work introduces {\it PrahokBART}, a compact pre-trained sequence-to-sequence model trained from scratch for Khmer using carefully curated Khmer and English corpora. We focus on improving the pre-training corpus quality and addressing the linguistic issues of Khmer, which are ignored in existing multilingual models, by incorporating linguistic components such as word segmentation and normalization. We evaluate PrahokBART on three generative tasks: machine translation, text summarization, and headline generation, where our results demonstrate that it outperforms mBART50, a strong multilingual pre-trained model. Additionally, our analysis provides insights into the impact of each linguistic module and evaluates how effectively our model handles space during text generation, which is crucial for the naturalness of texts in Khmer.

Shohei Higashiyama, Masao Ideuchi, Masao Utiyama

Entity linking is the task of associating linguistic expressions with entries in a knowledge base that represent real-world entities and concepts. Language resources for this task have primarily been developed for English, and the resources available for evaluating Japanese systems remain limited. In this study, we develop a corpus design policy for the entity linking task and construct an annotated corpus for training and evaluating Japanese entity linking systems, with rich coverage of linguistic expressions referring to entities that are specific to Japan. Evaluation of inter-annotator agreement confirms the high consistency of the annotations in the corpus, and a preliminary experiment on entity disambiguation based on string matching suggests that the corpus contains a substantial number of non-trivial cases, supporting its potential usefulness as an evaluation benchmark.

Jonas Belouadi, Eddy Ilg, Margret Keuper et al.

Automatically synthesizing figures from text captions is a compelling capability. However, achieving high geometric precision and editability requires representing figures as graphics programs in languages like TikZ, and aligned training data (i.e., graphics programs with captions) remains scarce. Meanwhile, large amounts of unaligned graphics programs and captioned raster images are more readily available. We reconcile these disparate data sources by presenting TikZero, which decouples graphics program generation from text understanding by using image representations as an intermediary bridge. It enables independent training on graphics programs and captioned images and allows for zero-shot text-guided graphics program synthesis during inference. We show that our method substantially outperforms baselines that can only operate with caption-aligned graphics programs. Furthermore, when leveraging caption-aligned graphics programs as a complementary training signal, TikZero matches or exceeds the performance of much larger models, including commercial systems like GPT-4o. Our code, datasets, and select models are publicly available.

Shohei Higashiyama, Hiroki Ouchi, Atsushi Fujita et al.

Geographic text, or textual data rich in geographic (geo-) information is a valuable source for various geographic applications, e.g., tourism management. Making such information accessible to speakers of other languages further enhances its utility; thus, accurate machine translation (MT) is essential for equity in multilingual geo-information access. To facilitate in-depth analysis for geographic text, we introduce ATD-Trans, a geographically grounded Japanese--English travelogue translation dataset, which enables evaluation of MT quality at both the overall and geo-entity levels across domestic (within Japan) and overseas regions. Our experiments on existing language models examine two factors: model language focus and geographic regions. The results highlight advantages of Japanese-enhanced models and greater difficulty in translating domestic-region geo-entities mentioned in travel blogs.

Haiyue Song, Johannes Eschbach-Dymanus, Hour Kaing et al.

Recent works on structured text translation remain limited to the sentence level, as they struggle to effectively handle the complex document-level XML or HTML structures. To address this, we propose \textbf{Format Reinforcement Learning (FormatRL)}, which employs Group Relative Policy Optimization on top of a supervised fine-tuning model to directly optimize novel structure-aware rewards: 1) TreeSim, which measures structural similarity between predicted and reference XML trees and 2) Node-chrF, which measures translation quality at the level of XML nodes. Additionally, we apply StrucAUC, a fine-grained metric distinguishing between minor errors and major structural failures. Experiments on the SAP software-documentation benchmark demonstrate improvements across six metrics and an analysis further shows how different reward functions contribute to improvements in both structural and translation quality.

Boxuan Lyu, Haiyue Song, Hidetaka Kamigaito et al.

Error Span Detection (ESD) is a subtask of automatic machine translation evaluation that localizes error spans in translations and labels their severity. State-of-the-art generative ESD methods typically decode using Maximum a Posteriori (MAP), assuming that model-estimated probabilities are perfectly correlated with similarity to human annotation. However, we observed that annotations dissimilar to the human annotation could achieve a higher model likelihood than the human annotation. We address this issue by applying Minimum Bayes Risk (MBR) decoding to generative ESD models. Specifically, we employ sentence- and span-level similarity metrics as utility functions to select candidate hypotheses based on their approximate similarity to the human annotation. Extensive experimental results show that our MBR decoding outperforms the MAP baseline at the system, sentence, and span-levels. Furthermore, to mitigate the computational cost of MBR decoding, we demonstrate that applying MBR distillation enables a standard greedy model to match MBR decoding performance, effectively eliminating the inference-time latency bottleneck.

Haiyue Song, Masao Utiyama

Continual pre-training is widely used to adapt LLMs to target languages and domains, yet the mixture ratio of training data remains a sensitive hyperparameter that is expensive to tune: they must be fixed before training begins, and a suboptimal choice can waste weeks of compute. In this work, we propose OptiMer, which decouples ratio selection from training: we train one CPT model per dataset, extract each model's distribution vector, which represents the parameter shift induced by that dataset, and search for optimal composition weights post-hoc via Bayesian optimization. Experiments on Gemma 3 27B across languages (Japanese, Chinese) and domains (Math, Code) show that OptiMer consistently outperforms data mixture and model averaging baselines with 15-35 times lower search cost. Key findings reveal that 1) the optimized weights can be interpreted as data mixture ratios, and retraining with these ratios improves data mixture CPT, and 2) the same vector pool can be re-optimized for a given objective without any retraining, producing target-tailored models on demand. Our work establishes that data mixture ratio selection, traditionally a pre-training decision, can be reformulated as a post-hoc optimization over distribution vectors, offering a more flexible paradigm for continual pre-training.

Zhi Qu, Yiran Wang, Jiannan Mao et al.

The multilingual neural machine translation (MNMT) aims for arbitrary translations across multiple languages. Although MNMT-specific models trained on parallel data offer low costs in training and deployment, their performance consistently lags behind that of large language models (LLMs). In this work, we introduce registering, a novel method that enables a small MNMT-specific model to compete with LLMs. Specifically, we insert a set of artificial tokens specifying the target language, called registers, into the input sequence between the source and target tokens. By modifying the attention mask, the target token generation only pays attention to the activation of registers, representing the source tokens in the target language space. Experiments on EC-40, a large-scale benchmark, show that our method advances the state-of-the-art of MNMT. We further pre-train two models, namely MITRE (multilingual translation with registers), by 9.3 billion sentence pairs across 24 languages collected from public corpora. One of them, MITRE-913M, outperforms NLLB-3.3B, achieves comparable performance with commercial LLMs, and shows strong adaptability in fine-tuning. Finally, we open-source our models to facilitate further research and development in MNMT: https://github.com/zhiqu22/mitre.

Xiaolin Wang, Masao Utiyama, Eiichiro Sumita

This paper presents an open-source neural machine translation toolkit named CytonMT (https://github.com/arthurxlw/cytonMt). The toolkit is built from scratch only using C++ and NVIDIA's GPU-accelerated libraries. The toolkit features training efficiency, code simplicity and translation quality. Benchmarks show that CytonMT accelerates the training speed by 64.5% to 110.8% on neural networks of various sizes, and achieves competitive translation quality.

Michael Paul, Kenji Imamura, Xiaolin Wang et al.

In line with the Global Communication Plan 2025, we have pursued the development of multilingual translation technologies to realize a language-barrier-free experience at Expo 2025 Osaka. Our work includes the advancement of simultaneous interpretation systems emphasizing high translation quality and low latency. Key achievements include chunk-based input segmentation, context-aware translation, and multi-engine machine translation technologies. Through demonstration deployments and collaboration with private companies, our technologies have led to real-world applications, with several services and systems showcased at Expo 2025 Osaka.

Hiroyuki Deguchi, Yusuke Sakai, Hidetaka Kamigaito et al.

Minimum Bayes risk (MBR) decoding achieved state-of-the-art translation performance by using COMET, a neural metric that has a high correlation with human evaluation. However, MBR decoding requires quadratic time since it computes the expected score between a translation hypothesis and all reference translations. We propose centroid-based MBR (CBMBR) decoding to improve the speed of MBR decoding. Our method clusters the reference translations in the feature space, and then calculates the score using the centroids of each cluster. The experimental results show that our CBMBR not only improved the decoding speed of the expected score calculation 5.7 times, but also outperformed vanilla MBR decoding in translation quality by up to 0.5 COMET in the WMT'22 En$\leftrightarrow$Ja, En$\leftrightarrow$De, En$\leftrightarrow$Zh, and WMT'23 En$\leftrightarrow$Ja translation tasks.

Haryo Akbarianto Wibowo, Haiyue Song, Hideki Tanaka et al.

Large Language Models (LLMs) have grown increasingly expensive to deploy, driving the need for effective model compression techniques. While block pruning offers a straightforward approach to reducing model size, existing methods often struggle to maintain performance or require substantial computational resources for recovery. We present IteRABRe, a simple yet effective iterative pruning method that achieves superior compression results while requiring minimal computational resources. Using only 2.5M tokens for recovery, our method outperforms baseline approaches by ~3% on average when compressing the Llama3.1-8B and Qwen2.5-7B models. IteRABRe demonstrates particular strength in the preservation of linguistic capabilities, showing an improvement 5% over the baselines in language-related tasks. Our analysis reveals distinct pruning characteristics between these models, while also demonstrating preservation of multilingual capabilities.

Ahmed Elshabrawy, Hour Kaing, Haiyue Song et al.

Alignment with high-resource standard languages is often assumed to aid the modeling of related low-resource varieties. We challenge this assumption by demonstrating that excessive representational entanglement with a dominant variety, such as Modern Standard Arabic (MSA) in relation to Arabic dialects, can actively hinder generative modeling. We present the first comprehensive causal study of this phenomenon by analyzing and directly intervening in the internal representation geometry of large language models (LLMs). Our key contribution is an online variational probing framework that continuously estimates the subspace of the standard variety during fine-tuning, enabling projection-based decoupling from this space. While our study uses Arabic as a case due to its unusually rich parallel resources across 25 dialects, the broader motivation is methodological: dialectal MT serves as a controlled proxy for generative tasks where comparable multi-variety corpora are unavailable. Across 25 dialects, our intervention improves generation quality by up to +4.9 chrF++ and +2.0 on average compared to standard fine-tuning, despite a measured tradeoff in standard-language performance. These results provide causal evidence that subspace dominance by high-resource varieties can restrict generative capacity for related varieties. More generally, we unify geometric and information-theoretic probing with subspace-level causal interventions, offering practical tools for improving generative modeling in closely related language families and, more broadly, for controlling representational allocation in multilingual and multi-domain LLMs. Code will be released.

Shohei Higashiyama, Masao Utiyama

Lexical normalization research has sought to tackle the challenge of processing informal expressions in user-generated text, yet the absence of comprehensive evaluations leaves it unclear which methods excel across multiple perspectives. Focusing on unsegmented languages, we make three key contributions: (1) creating a large-scale, multi-domain Japanese normalization dataset, (2) developing normalization methods based on state-of-the-art pretrained models, and (3) conducting experiments across multiple evaluation perspectives. Our experiments show that both encoder-only and decoder-only approaches achieve promising results in both accuracy and efficiency.

Zhi Qu, Yiran Wang, Chenchen Ding et al.

Existing multilingual neural machine translation (MNMT) approaches mainly focus on improving models with the encoder-decoder architecture to translate multiple languages. However, decoder-only architecture has been explored less in MNMT due to its underperformance when trained on parallel data solely. In this work, we attribute the issue of the decoder-only architecture to its lack of language transfer capability. Specifically, the decoder-only architecture is insufficient in encoding source tokens with the target language features. We propose dividing the decoding process into two stages so that target tokens are explicitly excluded in the first stage to implicitly boost the transfer capability across languages. Additionally, we impose contrastive learning on translation instructions, resulting in improved performance in zero-shot translation. We conduct experiments on TED-19 and OPUS-100 datasets, considering both training from scratch and fine-tuning scenarios. Experimental results show that, compared to the encoder-decoder architecture, our methods not only perform competitively in supervised translations but also achieve improvements of up to 3.39 BLEU, 6.99 chrF++, 3.22 BERTScore, and 4.81 COMET in zero-shot translations.

Yiran Wang, Masao Utiyama

Recently, binary representation has been proposed as a novel representation that lies between continuous and discrete representations. It exhibits considerable information-preserving capability when being used to replace continuous input vectors. In this paper, we investigate the feasibility of further introducing it to the output side, aiming to allow models to output binary labels instead. To preserve the structural information on the output side along with label information, we extend the previous contrastive hashing method as structured contrastive hashing. More specifically, we upgrade CKY from label-level to bit-level, define a new similarity function with span marginal probabilities, and introduce a novel contrastive loss function with a carefully designed instance selection strategy. Our model achieves competitive performance on various structured prediction tasks, and demonstrates that binary representation can be considered a novel representation that further bridges the gap between the continuous nature of deep learning and the discrete intrinsic property of natural languages.

Zhuosheng Zhang, Siru Ouyang, Hai Zhao et al.

Conversational machine reading (CMR) requires machines to communicate with humans through multi-turn interactions between two salient dialogue states of decision making and question generation processes. In open CMR settings, as the more realistic scenario, the retrieved background knowledge would be noisy, which results in severe challenges in the information transmission. Existing studies commonly train independent or pipeline systems for the two subtasks. However, those methods are trivial by using hard-label decisions to activate question generation, which eventually hinders the model performance. In this work, we propose an effective gating strategy by smoothing the two dialogue states in only one decoder and bridge decision making and question generation to provide a richer dialogue state reference. Experiments on the OR-ShARC dataset show the effectiveness of our method, which achieves new state-of-the-art results.

Zuchao Li, Kevin Parnow, Hai Zhao et al.

Though the pre-trained contextualized language model (PrLM) has made a significant impact on NLP, training PrLMs in languages other than English can be impractical for two reasons: other languages often lack corpora sufficient for training powerful PrLMs, and because of the commonalities among human languages, computationally expensive PrLM training for different languages is somewhat redundant. In this work, building upon the recent works connecting cross-lingual model transferring and neural machine translation, we thus propose a novel cross-lingual model transferring framework for PrLMs: TreLM. To handle the symbol order and sequence length differences between languages, we propose an intermediate ``TRILayer" structure that learns from these differences and creates a better transfer in our primary translation direction, as well as a new cross-lingual language modeling objective for transfer training. Additionally, we showcase an embedding aligning that adversarially adapts a PrLM's non-contextualized embedding space and the TRILayer structure to learn a text transformation network across languages, which addresses the vocabulary difference between languages. Experiments on both language understanding and structure parsing tasks show the proposed framework significantly outperforms language models trained from scratch with limited data in both performance and efficiency. Moreover, despite an insignificant performance loss compared to pre-training from scratch in resource-rich scenarios, our cross-lingual model transferring framework is significantly more economical.

Shohei Higashiyama, Masao Utiyama, Taro Watanabe et al.

Morphological analysis (MA) and lexical normalization (LN) are both important tasks for Japanese user-generated text (UGT). To evaluate and compare different MA/LN systems, we have constructed a publicly available Japanese UGT corpus. Our corpus comprises 929 sentences annotated with morphological and normalization information, along with category information we classified for frequent UGT-specific phenomena. Experiments on the corpus demonstrated the low performance of existing MA/LN methods for non-general words and non-standard forms, indicating that the corpus would be a challenging benchmark for further research on UGT.

Zuchao Li, Zhuosheng Zhang, Hai Zhao et al.

Text encoding is one of the most important steps in Natural Language Processing (NLP). It has been done well by the self-attention mechanism in the current state-of-the-art Transformer encoder, which has brought about significant improvements in the performance of many NLP tasks. Though the Transformer encoder may effectively capture general information in its resulting representations, the backbone information, meaning the gist of the input text, is not specifically focused on. In this paper, we propose explicit and implicit text compression approaches to enhance the Transformer encoding and evaluate models using this approach on several typical downstream tasks that rely on the encoding heavily. Our explicit text compression approaches use dedicated models to compress text, while our implicit text compression approach simply adds an additional module to the main model to handle text compression. We propose three ways of integration, namely backbone source-side fusion, target-side fusion, and both-side fusion, to integrate the backbone information into Transformer-based models for various downstream tasks. Our evaluation on benchmark datasets shows that the proposed explicit and implicit text compression approaches improve results in comparison to strong baselines. We therefore conclude, when comparing the encodings to the baseline models, text compression helps the encoders to learn better language representations.

Zhuosheng Zhang, Haojie Yu, Hai Zhao et al.

Word representation is a fundamental component in neural language understanding models. Recently, pre-trained language models (PrLMs) offer a new performant method of contextualized word representations by leveraging the sequence-level context for modeling. Although the PrLMs generally give more accurate contextualized word representations than non-contextualized models do, they are still subject to a sequence of text contexts without diverse hints for word representation from multimodality. This paper thus proposes a visual representation method to explicitly enhance conventional word embedding with multiple-aspect senses from visual guidance. In detail, we build a small-scale word-image dictionary from a multimodal seed dataset where each word corresponds to diverse related images. The texts and paired images are encoded in parallel, followed by an attention layer to integrate the multimodal representations. We show that the method substantially improves the accuracy of disambiguation. Experiments on 12 natural language understanding and machine translation tasks further verify the effectiveness and the generalization capability of the proposed approach.

Zuchao Li, Hai Zhao, Rui Wang et al.

In this paper, we introduced our joint team SJTU-NICT 's participation in the WMT 2020 machine translation shared task. In this shared task, we participated in four translation directions of three language pairs: English-Chinese, English-Polish on supervised machine translation track, German-Upper Sorbian on low-resource and unsupervised machine translation tracks. Based on different conditions of language pairs, we have experimented with diverse neural machine translation (NMT) techniques: document-enhanced NMT, XLM pre-trained language model enhanced NMT, bidirectional translation as a pre-training, reference language based UNMT, data-dependent gaussian prior objective, and BT-BLEU collaborative filtering self-training. We also used the TF-IDF algorithm to filter the training set to obtain a domain more similar set with the test set for finetuning. In our submissions, the primary systems won the first place on English to Chinese, Polish to English, and German to Upper Sorbian translation directions.

Akiko Aizawa, Frederic Bergeron, Junjie Chen et al.

The global pandemic of COVID-19 has made the public pay close attention to related news, covering various domains, such as sanitation, treatment, and effects on education. Meanwhile, the COVID-19 condition is very different among the countries (e.g., policies and development of the epidemic), and thus citizens would be interested in news in foreign countries. We build a system for worldwide COVID-19 information aggregation containing reliable articles from 10 regions in 7 languages sorted by topics. Our reliable COVID-19 related website dataset collected through crowdsourcing ensures the quality of the articles. A neural machine translation module translates articles in other languages into Japanese and English. A BERT-based topic-classifier trained on our article-topic pair dataset helps users find their interested information efficiently by putting articles into different categories.

Haipeng Sun, Rui Wang, Kehai Chen et al.

Unsupervised neural machine translation (UNMT) has recently achieved remarkable results for several language pairs. However, it can only translate between a single language pair and cannot produce translation results for multiple language pairs at the same time. That is, research on multilingual UNMT has been limited. In this paper, we empirically introduce a simple method to translate between thirteen languages using a single encoder and a single decoder, making use of multilingual data to improve UNMT for all language pairs. On the basis of the empirical findings, we propose two knowledge distillation methods to further enhance multilingual UNMT performance. Our experiments on a dataset with English translated to and from twelve other languages (including three language families and six language branches) show remarkable results, surpassing strong unsupervised individual baselines while achieving promising performance between non-English language pairs in zero-shot translation scenarios and alleviating poor performance in low-resource language pairs.

Haipeng Sun, Rui Wang, Kehai Chen et al.

Unsupervised neural machine translation (UNMT) that relies solely on massive monolingual corpora has achieved remarkable results in several translation tasks. However, in real-world scenarios, massive monolingual corpora do not exist for some extremely low-resource languages such as Estonian, and UNMT systems usually perform poorly when there is not adequate training corpus for one language. In this paper, we first define and analyze the unbalanced training data scenario for UNMT. Based on this scenario, we propose UNMT self-training mechanisms to train a robust UNMT system and improve its performance in this case. Experimental results on several language pairs show that the proposed methods substantially outperform conventional UNMT systems.

Kehai Chen, Rui Wang, Masao Utiyama et al.

In Transformer-based neural machine translation (NMT), the positional encoding mechanism helps the self-attention networks to learn the source representation with order dependency, which makes the Transformer-based NMT achieve state-of-the-art results for various translation tasks. However, Transformer-based NMT only adds representations of positions sequentially to word vectors in the input sentence and does not explicitly consider reordering information in this sentence. In this paper, we first empirically investigate the relationship between source reordering information and translation performance. The empirical findings show that the source input with the target order learned from the bilingual parallel dataset can substantially improve translation performance. Thus, we propose a novel reordering method to explicitly model this reordering information for the Transformer-based NMT. The empirical results on the WMT14 English-to-German, WAT ASPEC Japanese-to-English, and WMT17 Chinese-to-English translation tasks show the effectiveness of the proposed approach.

Zuchao Li, Hai Zhao, Rui Wang et al.

Exploiting a common language as an auxiliary for better translation has a long tradition in machine translation and lets supervised learning-based machine translation enjoy the enhancement delivered by the well-used pivot language in the absence of a source language to target language parallel corpus. The rise of unsupervised neural machine translation (UNMT) almost completely relieves the parallel corpus curse, though UNMT is still subject to unsatisfactory performance due to the vagueness of the clues available for its core back-translation training. Further enriching the idea of pivot translation by extending the use of parallel corpora beyond the source-target paradigm, we propose a new reference language-based framework for UNMT, RUNMT, in which the reference language only shares a parallel corpus with the source, but this corpus still indicates a signal clear enough to help the reconstruction training of UNMT through a proposed reference agreement mechanism. Experimental results show that our methods improve the quality of UNMT over that of a strong baseline that uses only one auxiliary language, demonstrating the usefulness of the proposed reference language-based UNMT and establishing a good start for the community.

Chaoqun Duan, Kehai Chen, Rui Wang et al.

Existing neural machine translation (NMT) systems utilize sequence-to-sequence neural networks to generate target translation word by word, and then make the generated word at each time-step and the counterpart in the references as consistent as possible. However, the trained translation model tends to focus on ensuring the accuracy of the generated target word at the current time-step and does not consider its future cost which means the expected cost of generating the subsequent target translation (i.e., the next target word). To respond to this issue, we propose a simple and effective method to model the future cost of each target word for NMT systems. In detail, a time-dependent future cost is estimated based on the current generated target word and its contextual information to boost the training of the NMT model. Furthermore, the learned future context representation at the current time-step is used to help the generation of the next target word in the decoding. Experimental results on three widely-used translation datasets, including the WMT14 German-to-English, WMT14 English-to-French, and WMT17 Chinese-to-English, show that the proposed approach achieves significant improvements over strong Transformer-based NMT baseline.

Haipeng Sun, Rui Wang, Kehai Chen et al.

Unsupervised neural machine translation (UNMT) has recently attracted great interest in the machine translation community. The main advantage of the UNMT lies in its easy collection of required large training text sentences while with only a slightly worse performance than supervised neural machine translation which requires expensive annotated translation pairs on some translation tasks. In most studies, the UMNT is trained with clean data without considering its robustness to the noisy data. However, in real-world scenarios, there usually exists noise in the collected input sentences which degrades the performance of the translation system since the UNMT is sensitive to the small perturbations of the input sentences. In this paper, we first time explicitly take the noisy data into consideration to improve the robustness of the UNMT based systems. First of all, we clearly defined two types of noises in training sentences, i.e., word noise and word order noise, and empirically investigate its effect in the UNMT, then we propose adversarial training methods with denoising process in the UNMT. Experimental results on several language pairs show that our proposed methods substantially improved the robustness of the conventional UNMT systems in noisy scenarios.

Zuchao Li, Rui Wang, Kehai Chen et al.

State-of-the-art Transformer-based neural machine translation (NMT) systems still follow a standard encoder-decoder framework, in which source sentence representation can be well done by an encoder with self-attention mechanism. Though Transformer-based encoder may effectively capture general information in its resulting source sentence representation, the backbone information, which stands for the gist of a sentence, is not specifically focused on. In this paper, we propose an explicit sentence compression method to enhance the source sentence representation for NMT. In practice, an explicit sentence compression goal used to learn the backbone information in a sentence. We propose three ways, including backbone source-side fusion, target-side fusion, and both-side fusion, to integrate the compressed sentence into NMT. Our empirical tests on the WMT English-to-French and English-to-German translation tasks show that the proposed sentence compression method significantly improves the translation performances over strong baselines.

Zhuosheng Zhang, Rui Wang, Kehai Chen et al.

We present a universal framework to model contextualized sentence representations with visual awareness that is motivated to overcome the shortcomings of the multimodal parallel data with manual annotations. For each sentence, we first retrieve a diversity of images from a shared cross-modal embedding space, which is pre-trained on a large-scale of text-image pairs. Then, the texts and images are respectively encoded by transformer encoder and convolutional neural network. The two sequences of representations are further fused by a simple and effective attention layer. The architecture can be easily applied to text-only natural language processing tasks without manually annotating multimodal parallel corpora. We apply the proposed method on three tasks, including neural machine translation, natural language inference and sequence labeling and experimental results verify the effectiveness.

Shu Jiang, Rui Wang, Zuchao Li et al.

Standard neural machine translation (NMT) is on the assumption that the document-level context is independent. Most existing document-level NMT approaches are satisfied with a smattering sense of global document-level information, while this work focuses on exploiting detailed document-level context in terms of a memory network. The capacity of the memory network that detecting the most relevant part of the current sentence from memory renders a natural solution to model the rich document-level context. In this work, the proposed document-aware memory network is implemented to enhance the Transformer NMT baseline. Experiments on several tasks show that the proposed method significantly improves the NMT performance over strong Transformer baselines and other related studies.

Junya Ono, Masao Utiyama, Eiichiro Sumita

Reduction of training time is an important issue in many tasks like patent translation involving neural networks. Data parallelism and model parallelism are two common approaches for reducing training time using multiple graphics processing units (GPUs) on one machine. In this paper, we propose a hybrid data-model parallel approach for sequence-to-sequence (Seq2Seq) recurrent neural network (RNN) machine translation. We apply a model parallel approach to the RNN encoder-decoder part of the Seq2Seq model and a data parallel approach to the attention-softmax part of the model. We achieved a speed-up of 4.13 to 4.20 times when using 4 GPUs compared with the training speed when using 1 GPU without affecting machine translation accuracy as measured in terms of BLEU scores.

Haipeng Sun, Rui Wang, Kehai Chen et al.

Domain adaptation has been well-studied in supervised neural machine translation (SNMT). However, it has not been well-studied for unsupervised neural machine translation (UNMT), although UNMT has recently achieved remarkable results in several domain-specific language pairs. Besides the inconsistent domains between training data and test data for SNMT, there sometimes exists an inconsistent domain between two monolingual training data for UNMT. In this work, we empirically show different scenarios for unsupervised neural machine translation. Based on these scenarios, we revisit the effect of the existing domain adaptation methods including batch weighting and fine tuning methods in UNMT. Finally, we propose modified methods to improve the performances of domain-specific UNMT systems.

Rui Wang, Benjamin Marie, Masao Utiyama et al.

This paper presents the NICT's participation in the WMT18 shared parallel corpus filtering task. The organizers provided 1 billion words German-English corpus crawled from the web as part of the Paracrawl project. This corpus is too noisy to build an acceptable neural machine translation (NMT) system. Using the clean data of the WMT18 shared news translation task, we designed several features and trained a classifier to score each sentence pairs in the noisy data. Finally, we sampled 100 million and 10 million words and built corresponding NMT systems. Empirical results show that our NMT systems trained on sampled data achieve promising performance.

Benjamin Marie, Rui Wang, Atsushi Fujita et al.

This paper presents the NICT's participation to the WMT18 shared news translation task. We participated in the eight translation directions of four language pairs: Estonian-English, Finnish-English, Turkish-English and Chinese-English. For each translation direction, we prepared state-of-the-art statistical (SMT) and neural (NMT) machine translation systems. Our NMT systems were trained with the transformer architecture using the provided parallel data enlarged with a large quantity of back-translated monolingual data that we generated with a new incremental training framework. Our primary submissions to the task are the result of a simple combination of our SMT and NMT systems. Our systems are ranked first for the Estonian-English and Finnish-English language pairs (constraint) according to BLEU-cased.

Zhisong Zhang, Rui Wang, Masao Utiyama et al.

In Neural Machine Translation (NMT), the decoder can capture the features of the entire prediction history with neural connections and representations. This means that partial hypotheses with different prefixes will be regarded differently no matter how similar they are. However, this might be inefficient since some partial hypotheses can contain only local differences that will not influence future predictions. In this work, we introduce recombination in NMT decoding based on the concept of the "equivalence" of partial hypotheses. Heuristically, we use a simple $n$-gram suffix based equivalence function and adapt it into beam search decoding. Through experiments on large-scale Chinese-to-English and English-to-Germen translation tasks, we show that the proposed method can obtain similar translation quality with a smaller beam size, making NMT decoding more efficient.

Rui Wang, Masao Utiyama, Eiichiro Sumita

Traditional Neural machine translation (NMT) involves a fixed training procedure where each sentence is sampled once during each epoch. In reality, some sentences are well-learned during the initial few epochs; however, using this approach, the well-learned sentences would continue to be trained along with those sentences that were not well learned for 10-30 epochs, which results in a wastage of time. Here, we propose an efficient method to dynamically sample the sentences in order to accelerate the NMT training. In this approach, a weight is assigned to each sentence based on the measured difference between the training costs of two iterations. Further, in each epoch, a certain percentage of sentences are dynamically sampled according to their weights. Empirical results based on the NIST Chinese-to-English and the WMT English-to-German tasks depict that the proposed method can significantly accelerate the NMT training and improve the NMT performance.

Jingyi Zhang, Masao Utiyama, Eiichro Sumita et al.

One of the difficulties of neural machine translation (NMT) is the recall and appropriate translation of low-frequency words or phrases. In this paper, we propose a simple, fast, and effective method for recalling previously seen translation examples and incorporating them into the NMT decoding process. Specifically, for an input sentence, we use a search engine to retrieve sentence pairs whose source sides are similar with the input sentence, and then collect $n$-grams that are both in the retrieved target sentences and aligned with words that match in the source sentences, which we call "translation pieces". We compute pseudo-probabilities for each retrieved sentence based on similarities between the input sentence and the retrieved source sentences, and use these to weight the retrieved translation pieces. Finally, an existing NMT model is used to translate the input sentence, with an additional bonus given to outputs that contain the collected translation pieces. We show our method improves NMT translation results up to 6 BLEU points on three narrow domain translation tasks where repetitiveness of the target sentences is particularly salient. It also causes little increase in the translation time, and compares favorably to another alternative retrieval-based method with respect to accuracy, speed, and simplicity of implementation.

Kehai Chen, Rui Wang, Masao Utiyama et al.

Attention mechanism, including global attention and local attention, plays a key role in neural machine translation (NMT). Global attention attends to all source words for word prediction. In comparison, local attention selectively looks at fixed-window source words. However, alignment weights for the current target word often decrease to the left and right by linear distance centering on the aligned source position and neglect syntax-directed distance constraints. In this paper, we extend local attention with syntax-distance constraint, to focus on syntactically related source words with the predicted target word, thus learning a more effective context vector for word prediction. Moreover, we further propose a double context NMT architecture, which consists of a global context vector and a syntax-directed context vector over the global attention, to provide more translation performance for NMT from source representation. The experiments on the large-scale Chinese-to-English and English-to-Germen translation tasks show that the proposed approach achieves a substantial and significant improvement over the baseline system.

Jingyi Zhang, Masao Utiyama, Eiichro Sumita et al.

Compared to traditional statistical machine translation (SMT), neural machine translation (NMT) often sacrifices adequacy for the sake of fluency. We propose a method to combine the advantages of traditional SMT and NMT by exploiting an existing phrase-based SMT model to compute the phrase-based decoding cost for an NMT output and then using this cost to rerank the n-best NMT outputs. The main challenge in implementing this approach is that NMT outputs may not be in the search space of the standard phrase-based decoding algorithm, because the search space of phrase-based SMT is limited by the phrase-based translation rule table. We propose a soft forced decoding algorithm, which can always successfully find a decoding path for any NMT output. We show that using the forced decoding cost to rerank the NMT outputs can successfully improve translation quality on four different language pairs.

Lemao Liu, Masao Utiyama, Andrew Finch et al.

The attention mechanisim is appealing for neural machine translation, since it is able to dynam- ically encode a source sentence by generating a alignment between a target word and source words. Unfortunately, it has been proved to be worse than conventional alignment models in aligment accuracy. In this paper, we analyze and explain this issue from the point view of re- ordering, and propose a supervised attention which is learned with guidance from conventional alignment models. Experiments on two Chinese-to-English translation tasks show that the super- vised attention mechanism yields better alignments leading to substantial gains over the standard attention based NMT.

Rui Wang, Hai Zhao, Bao-Liang Lu et al.

Although more additional corpora are now available for Statistical Machine Translation (SMT), only the ones which belong to the same or similar domains with the original corpus can indeed enhance SMT performance directly. Most of the existing adaptation methods focus on sentence selection. In comparison, phrase is a smaller and more fine grained unit for data selection, therefore we propose a straightforward and efficient connecting phrase based adaptation method, which is applied to both bilingual phrase pair and monolingual n-gram adaptation. The proposed method is evaluated on IWSLT/NIST data sets, and the results show that phrase based SMT performance are significantly improved (up to +1.6 in comparison with phrase based SMT baseline system and +0.9 in comparison with existing methods).

Rui Wang, Hai Zhao, Sabine Ploux et al.

Most of the existing methods for bilingual word embedding only consider shallow context or simple co-occurrence information. In this paper, we propose a latent bilingual sense unit (Bilingual Sense Clique, BSC), which is derived from a maximum complete sub-graph of pointwise mutual information based graph over bilingual corpus. In this way, we treat source and target words equally and a separated bilingual projection processing that have to be used in most existing works is not necessary any more. Several dimension reduction methods are evaluated to summarize the BSC-word relationship. The proposed method is evaluated on bilingual lexicon translation tasks and empirical results show that bilingual sense embedding methods outperform existing bilingual word embedding methods.