In a practical scenario, multi-domain neural machine translation (MDNMT) aims to continuously acquire knowledge from new domain data while retaining old knowledge. Previous work separately learns each new domain knowledge based on parameter isolation methods, which effectively capture the new knowledge. However, task-specific parameters lead to isolation between models, which hinders the mutual transfer of knowledge between new domains. Given the scarcity of domain-specific corpora, we consider making full use of the data from multiple new domains. Therefore, our work aims to leverage previously acquired domain knowledge when modeling subsequent domains. To this end, we propose an Iterative Continual Learning (ICL) framework for multi-domain neural machine translation. Specifically, when each new domain arrives, (1) we first build a pluggable incremental learning model, (2) then we design an iterative updating algorithm to continuously update the original model, which can be used flexibly for constructing subsequent domain models. Furthermore, we design a domain knowledge transfer mechanism to enhance the fine-grained domain-specific representation, thereby solving the word ambiguity caused by mixing domain data. Experimental results on the UM-Corpus and OPUS multi-domain datasets show the superior performance of our proposed model compared to representative baselines.
Currently, multi-domain neural machine translation (NMT) has become a significant research topic in domain adaptation machine translation, which trains a single model by mixing data from multiple domains. Multi-domain NMT aims to improve the performance of the low-resources domain through data augmentation. However, mixed domain data brings more translation ambiguity. Previous work focused on domain-general or domain-context knowledge learning, respectively. Therefore, there is a challenge for acquiring domain-general or domain-context knowledge simultaneously. To this end, we propose a unified framework for learning simultaneously domain-general and domain-specific knowledge, we are the first to apply parameter differentiation in multi-domain NMT. Specifically, we design the differentiation criterion and differentiation granularity to obtain domain-specific parameters. Experimental results on multi-domain UM-corpus English-to-Chinese and OPUS German-to-English datasets show that the average BLEU scores of the proposed method exceed the strong baseline by 1.22 and 1.87, respectively. In addition, we investigate the case study to illustrate the effectiveness of the proposed method in acquiring domain knowledge.
Generating coherent long texts is an important yet challenging task, particularly forthe open-ended generation. Prior work based on discrete latent codes focuses on the modeling of discourse relation, resulting in discrete codes only learning shallow semantics (Ji and Huang, 2021). A natural text always revolves around several related topics and the transition across them is natural and smooth.In this work, we investigate whether discrete latent codes can learn information of topics. To this end, we build a topic-aware latent code-guided text generation model. To encourage discrete codes to model information about topics, we propose a span-level bag-of-words training objective for the model. Automatic and manual evaluation experiments show that our method can generate more topic-relevant and coherent texts.
Syntax-controlled paraphrase generation aims to produce paraphrase conform to given syntactic patterns. To address this task, recent works have started to use parse trees (or syntactic templates) to guide generation.A constituency parse tree contains abundant structural information, such as parent-child relation, sibling relation, and the alignment relation between words and nodes. Previous works have only utilized parent-child and alignment relations, which may affect the generation quality. To address this limitation, we propose a Structural Information-augmented Syntax-Controlled Paraphrasing (SI-SCP) model. Particularly, we design a syntax encoder based on tree-transformer to capture parent-child and sibling relations. To model the alignment relation between words and nodes, we propose an attention regularization objective, which makes the decoder accurately select corresponding syntax nodes to guide the generation of words. Experiments show that SI-SCP achieves state-of-the-art performances in terms of semantic and syntactic quality on two popular benchmark datasets. Additionally, we propose a Syntactic Template Retriever (STR) to retrieve compatible syntactic structures. We validate that STR is capable of retrieving compatible syntactic structures. We further demonstrate the effectiveness of SI-SCP to generate diverse paraphrases with retrieved syntactic structures.
Previous works on syntactically controlled paraphrase generation heavily rely on large-scale parallel paraphrase data that is not easily available for many languages and domains. In this paper, we take this research direction to the extreme and investigate whether it is possible to learn syntactically controlled paraphrase generation with nonparallel data. We propose a syntactically-informed unsupervised paraphrasing model based on conditional variational auto-encoder (VAE) which can generate texts in a specified syntactic structure. Particularly, we design a two-stage learning method to effectively train the model using non-parallel data. The conditional VAE is trained to reconstruct the input sentence according to the given input and its syntactic structure. Furthermore, to improve the syntactic controllability and semantic consistency of the pre-trained conditional VAE, we fine-tune it using syntax controlling and cycle reconstruction learning objectives, and employ Gumbel-Softmax to combine these new learning objectives. Experiment results demonstrate that the proposed model trained only on non-parallel data is capable of generating diverse paraphrases with specified syntactic structure. Additionally, we validate the effectiveness of our method for generating syntactically adversarial examples on the sentiment analysis task.
Paraphrase generation (PG) is of great importance to many downstream tasks in natural language processing. Diversity is an essential nature to PG for enhancing generalization capability and robustness of downstream applications. Recently, neural sequence-to-sequence (Seq2Seq) models have shown promising results in PG. However, traditional model training for PG focuses on optimizing model prediction against single reference and employs cross-entropy loss, which objective is unable to encourage model to generate diverse paraphrases. In this work, we present a novel approach with multi-objective learning to PG. We propose a learning-exploring method to generate sentences as learning objectives from the learned data distribution, and employ reinforcement learning to combine these new learning objectives for model training. We first design a sample-based algorithm to explore diverse sentences. Then we introduce several reward functions to evaluate the sampled sentences as learning signals in terms of expressive diversity and semantic fidelity, aiming to generate diverse and high-quality paraphrases. To effectively optimize model performance satisfying different evaluating aspects, we use a GradNorm-based algorithm that automatically balances these training objectives. Experiments and analyses on Quora and Twitter datasets demonstrate that our proposed method not only gains a significant increase in diversity but also improves generation quality over several state-of-the-art baselines.
In Chinese dependency parsing, the joint model of word segmentation, POS tagging and dependency parsing has become the mainstream framework because it can eliminate error propagation and share knowledge, where the transition-based model with feature templates maintains the best performance. Recently, the graph-based joint model (Yan et al., 2019) on word segmentation and dependency parsing has achieved better performance, demonstrating the advantages of the graph-based models. However, this work can not provide POS information for downstream tasks, and the POS tagging task was proved to be helpful to the dependency parsing according to the research of the transition-based model. Therefore, we propose a graph-based joint model for Chinese word segmentation, POS tagging and dependency parsing. We designed a charater-level POS tagging task, and then train it jointly with the model of Yan et al. (2019). We adopt two methods of joint POS tagging task, one is by sharing parameters, the other is by using tag attention mechanism, which enables the three tasks to better share intermediate information and improve each other’s performance. The experimental results on the Penn Chinese treebank (CTB5) show that our proposed joint model improved by 0.38% on dependency parsing than the model of Yan et al. (2019). Compared with the best transition-based joint model, our model improved by 0.18%, 0.35% and 5.99% respectively in terms of word segmentation, POS tagging and dependency parsing.
Sentence matching is a key issue in natural language inference and paraphrase identification. Despite the recent progress on multi-layered neural network with cross sentence attention, one sentence learns attention to the intermediate representations of another sentence, which are propagated from preceding layers and therefore are uncertain and unstable for matching, particularly at the risk of error propagation. In this paper, we present an original semantics-oriented attention and deep fusion network (OSOA-DFN) for sentence matching. Unlike existing models, each attention layer of OSOA-DFN is oriented to the original semantic representation of another sentence, which captures the relevant information from a fixed matching target. The multiple attention layers allow one sentence to repeatedly read the important information of another sentence for better matching. We then additionally design deep fusion to propagate the attention information at each matching layer. At last, we introduce a self-attention mechanism to capture global context to enhance attention-aware representation within each sentence. Experiment results on three sentence matching benchmark datasets SNLI, SciTail and Quora show that OSOA-DFN has the ability to model sentence matching more precisely.
This paper presents our machine translation system that developed for the WAT2016 evalua-tion tasks of ja-en, ja-zh, en-ja, zh-ja, JPCja-en, JPCja-zh, JPCen-ja, JPCzh-ja. We build our system based on encoder–decoder framework by integrating recurrent neural network (RNN) and gate recurrent unit (GRU), and we also adopt an attention mechanism for solving the problem of information loss. Additionally, we propose a simple translation-specific approach to resolve the unknown word translation problem. Experimental results show that our system performs better than the baseline statistical machine translation (SMT) systems in each task. Moreover, it shows that our proposed approach of unknown word translation performs effec-tively improvement of translation results.
Parallel corpora are critical resources for machine translation research and development since parallel corpora contain translation equivalences of various granularities. Manual annotation of word & phrase alignment is of significance to provide gold-standard for developing and evaluating both example-based machine translation model and statistical machine translation model. This paper presents the work of word & phrase alignment annotation in the NICT Japanese-Chinese parallel corpus, which is constructed at the National Institute of Information and Communications Technology (NICT). We describe the specification of word alignment annotation and the tools specially developed for the manual annotation. The manual annotation on 17,000 sentence pairs has been completed. We examined the manually annotated word alignment data and extracted translation knowledge from the word & phrase aligned corpus.
Since 1994, China’s HTRDP machine translation evaluation has been conducted for five times. Systems of various translation directions between Chinese, English, Japanese and French have been tested. Both human evaluation and automatic evaluation are conducted in HTRDP evaluation. In recent years, the evaluation was organized jointly with NICT of Japan. This paper introduces some details of this evaluation.
We are constricting a Japanese-Chinese parallel corpus, which is a part of the NICT Multilingual Corpora. The corpus is general domain, of large scale of about 40,000 sentence pairs, long sentences, annotated with detailed information and high quality. To the best of our knowledge, this will be the first annotated Japanese-Chinese parallel corpus in the world. We created the corpus by selecting Japanese sentences from Mainichi Newspaper and then manually translating them into Chinese. We then annotated the corpus with morphological and syntactic structures and alignments at word and phrase levels. This paper describes the specification in human translation and detailed information annotation, and the tools we developed in the project. The experience we obtained and points we paid special attentions are also introduced for share with other researches in corpora construction.
Automatic word alignment is an important technology for extracting translation knowledge from parallel corpora. However, automatic techniques cannot resolve this problem completely because of variances in translations. We therefore need to investigate the performance potential of automatic word alignment and then decide how to suitably apply it. In this paper we first propose a lexical knowledge-based approach to word alignment on a Japanese-Chinese corpus. Then we evaluate the performance of the proposed approach on the corpus. At the same time we also apply a statistics-based approach, the well-known toolkit GIZA++, to the same test data. Through comparison of the performances of the two approaches, we propose a multi-aligner, exploiting the lexical knowledge-based aligner and the statistics-based aligner at the same time. Quantitative results confirmed the effectiveness of the multi-aligner.