ACL 2020 Papers

This paper presents an investigation on the distribution of word vectors belonging to a certain word class in a pre-trained word vector space. To this end, we made several assumptions about the distribution, modeled the distribution accordingly, and validated each assumption by comparing the goodness of each model. Specifically, we considered two types of word classes – the semantic class of direct objects of a verb and the semantic class in a thesaurus – and tried to build models that properly estimate how likely it is that a word in the vector space is a member of a given word class. Our results on selectional preference and WordNet datasets show that the centroid-based model will fail to achieve good enough performance, the geometry of the distribution and the existence of subgroups will have limited impact, and also the negative instances need to be considered for adequate modeling of the distribution. We further investigated the relationship between the scores calculated by each model and the degree of membership and found that discriminative learning-based models are best in finding the boundaries of a class, while models based on the offset between positive and negative instances perform best in determining the degree of membership.

Hypernymy detection, a.k.a, lexical entailment, is a fundamental sub-task of many natural language understanding tasks. Previous explorations mostly focus on monolingual hypernymy detection on high-resource languages, e.g., English, but few investigate the low-resource scenarios. This paper addresses the problem of low-resource hypernymy detection by combining high-resource languages. We extensively compare three joint training paradigms and for the first time propose applying meta learning to relieve the low-resource issue. Experiments demonstrate the superiority of our method among the three settings, which substantially improves the performance of extremely low-resource languages by preventing over-fitting on small datasets.

Biomedical named entities often play important roles in many biomedical text mining tools. However, due to the incompleteness of provided synonyms and numerous variations in their surface forms, normalization of biomedical entities is very challenging. In this paper, we focus on learning representations of biomedical entities solely based on the synonyms of entities. To learn from the incomplete synonyms, we use a model-based candidate selection and maximize the marginal likelihood of the synonyms present in top candidates. Our model-based candidates are iteratively updated to contain more difficult negative samples as our model evolves. In this way, we avoid the explicit pre-selection of negative samples from more than 400K candidates. On four biomedical entity normalization datasets having three different entity types (disease, chemical, adverse reaction), our model BioSyn consistently outperforms previous state-of-the-art models almost reaching the upper bound on each dataset.

The hypernymy detection task has been addressed under various frameworks. Previously, the design of unsupervised hypernymy scores has been extensively studied. In contrast, supervised classifiers, especially distributional models, leverage the global contexts of terms to make predictions, but are more likely to suffer from “lexical memorization”. In this work, we revisit supervised distributional models for hypernymy detection. Rather than taking embeddings of two terms as classification inputs, we introduce a representation learning framework named Bidirectional Residual Relation Embeddings (BiRRE). In this model, a term pair is represented by a BiRRE vector as features for hypernymy classification, which models the possibility of a term being mapped to another in the embedding space by hypernymy relations. A Latent Projection Model with Negative Regularization (LPMNR) is proposed to simulate how hypernyms and hyponyms are generated by neural language models, and to generate BiRRE vectors based on bidirectional residuals of projections. Experiments verify BiRRE outperforms strong baselines over various evaluation frameworks.

Open-domain dialogue generation has gained increasing attention in Natural Language Processing. Its evaluation requires a holistic means. Human ratings are deemed as the gold standard. As human evaluation is inefficient and costly, an automated substitute is highly desirable. In this paper, we propose holistic evaluation metrics that capture different aspects of open-domain dialogues. Our metrics consist of (1) GPT-2 based context coherence between sentences in a dialogue, (2) GPT-2 based fluency in phrasing, (3) n-gram based diversity in responses to augmented queries, and (4) textual-entailment-inference based logical self-consistency. The empirical validity of our metrics is demonstrated by strong correlations with human judgments. We open source the code and relevant materials.

The recent proliferation of ”fake news” has triggered a number of responses, most notably the emergence of several manual fact-checking initiatives. As a result and over time, a large number of fact-checked claims have been accumulated, which increases the likelihood that a new claim in social media or a new statement by a politician might have already been fact-checked by some trusted fact-checking organization, as viral claims often come back after a while in social media, and politicians like to repeat their favorite statements, true or false, over and over again. As manual fact-checking is very time-consuming (and fully automatic fact-checking has credibility issues), it is important to try to save this effort and to avoid wasting time on claims that have already been fact-checked. Interestingly, despite the importance of the task, it has been largely ignored by the research community so far. Here, we aim to bridge this gap. In particular, we formulate the task and we discuss how it relates to, but also differs from, previous work. We further create a specialized dataset, which we release to the research community. Finally, we present learning-to-rank experiments that demonstrate sizable improvements over state-of-the-art retrieval and textual similarity approaches.

News recommendation is an important technique for personalized news service. Compared with product and movie recommendations which have been comprehensively studied, the research on news recommendation is much more limited, mainly due to the lack of a high-quality benchmark dataset. In this paper, we present a large-scale dataset named MIND for news recommendation. Constructed from the user click logs of Microsoft News, MIND contains 1 million users and more than 160k English news articles, each of which has rich textual content such as title, abstract and body. We demonstrate MIND a good testbed for news recommendation through a comparative study of several state-of-the-art news recommendation methods which are originally developed on different proprietary datasets. Our results show the performance of news recommendation highly relies on the quality of news content understanding and user interest modeling. Many natural language processing techniques such as effective text representation methods and pre-trained language models can effectively improve the performance of news recommendation. The MIND dataset will be available at https://msnews.github.io.

Recently, large-scale datasets have vastly facilitated the development in nearly all domains of Natural Language Processing. However, there is currently no cross-task dataset in NLP, which hinders the development of multi-task learning. We propose MATINF, the first jointly labeled large-scale dataset for classification, question answering and summarization. MATINF contains 1.07 million question-answer pairs with human-labeled categories and user-generated question descriptions. Based on such rich information, MATINF is applicable for three major NLP tasks, including classification, question answering, and summarization. We benchmark existing methods and a novel multi-task baseline over MATINF to inspire further research. Our comprehensive comparison and experiments over MATINF and other datasets demonstrate the merits held by MATINF.

Code-switching is the use of more than one language in the same conversation or utterance. Recently, multilingual contextual embedding models, trained on multiple monolingual corpora, have shown promising results on cross-lingual and multilingual tasks. We present an evaluation benchmark, GLUECoS, for code-switched languages, that spans several NLP tasks in English-Hindi and English-Spanish. Specifically, our evaluation benchmark includes Language Identification from text, POS tagging, Named Entity Recognition, Sentiment Analysis, Question Answering and a new task for code-switching, Natural Language Inference. We present results on all these tasks using cross-lingual word embedding models and multilingual models. In addition, we fine-tune multilingual models on artificially generated code-switched data. Although multilingual models perform significantly better than cross-lingual models, our results show that in most tasks, across both language pairs, multilingual models fine-tuned on code-switched data perform best, showing that multilingual models can be further optimized for code-switching tasks.

This paper presents the problem of conversational plotting agents that carry out plotting actions from natural language instructions. To facilitate the development of such agents, we introduce ChartDialogs, a new multi-turn dialog dataset, covering a popular plotting library, matplotlib. The dataset contains over 15,000 dialog turns from 3,200 dialogs covering the majority of matplotlib plot types. Extensive experiments show the best-performing method achieving 61% plotting accuracy, demonstrating that the dataset presents a non-trivial challenge for future research on this task.

We propose an automatic evaluation method of machine translation that uses source language sentences regarded as additional pseudo references. The proposed method evaluates a translation hypothesis in a regression model. The model takes the paired source, reference, and hypothesis sentence all together as an input. A pretrained large scale cross-lingual language model encodes the input to sentence-pair vectors, and the model predicts a human evaluation score with those vectors. Our experiments show that our proposed method using Cross-lingual Language Model (XLM) trained with a translation language modeling (TLM) objective achieves a higher correlation with human judgments than a baseline method that uses only hypothesis and reference sentences. Additionally, using source sentences in our proposed method is confirmed to improve the evaluation performance.

The standard training algorithm in neural machine translation (NMT) suffers from exposure bias, and alternative algorithms have been proposed to mitigate this. However, the practical impact of exposure bias is under debate. In this paper, we link exposure bias to another well-known problem in NMT, namely the tendency to generate hallucinations under domain shift. In experiments on three datasets with multiple test domains, we show that exposure bias is partially to blame for hallucinations, and that training with Minimum Risk Training, which avoids exposure bias, can mitigate this. Our analysis explains why exposure bias is more problematic under domain shift, and also links exposure bias to the beam search problem, i.e. performance deterioration with increasing beam size. Our results provide a new justification for methods that reduce exposure bias: even if they do not increase performance on in-domain test sets, they can increase model robustness to domain shift.

This paper proposes a simple and effective algorithm for incorporating lexical constraints in neural machine translation. Previous work either required re-training existing models with the lexical constraints or incorporating them during beam search decoding with significantly higher computational overheads. Leveraging the flexibility and speed of a recently proposed Levenshtein Transformer model (Gu et al., 2019), our method injects terminology constraints at inference time without any impact on decoding speed. Our method does not require any modification to the training procedure and can be easily applied at runtime with custom dictionaries. Experiments on English-German WMT datasets show that our approach improves an unconstrained baseline and previous approaches.

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.

The choice of hyper-parameters affects the performance of neural models. While much previous research (Sutskever et al., 2013; Duchi et al., 2011; Kingma and Ba, 2015) focuses on accelerating convergence and reducing the effects of the learning rate, comparatively few papers concentrate on the effect of batch size. In this paper, we analyze how increasing batch size affects gradient direction, and propose to evaluate the stability of gradients with their angle change. Based on our observations, the angle change of gradient direction first tends to stabilize (i.e. gradually decrease) while accumulating mini-batches, and then starts to fluctuate. We propose to automatically and dynamically determine batch sizes by accumulating gradients of mini-batches and performing an optimization step at just the time when the direction of gradients starts to fluctuate. To improve the efficiency of our approach for large models, we propose a sampling approach to select gradients of parameters sensitive to the batch size. Our approach dynamically determines proper and efficient batch sizes during training. In our experiments on the WMT 14 English to German and English to French tasks, our approach improves the Transformer with a fixed 25k batch size by +0.73 and +0.82 BLEU respectively.

In encoder-decoder neural models, multiple encoders are in general used to represent the contextual information in addition to the individual sentence. In this paper, we investigate multi-encoder approaches in document-level neural machine translation (NMT). Surprisingly, we find that the context encoder does not only encode the surrounding sentences but also behaves as a noise generator. This makes us rethink the real benefits of multi-encoder in context-aware translation - some of the improvements come from robust training. We compare several methods that introduce noise and/or well-tuned dropout setup into the training of these encoders. Experimental results show that noisy training plays an important role in multi-encoder-based NMT, especially when the training data is small. Also, we establish a new state-of-the-art on IWSLT Fr-En task by careful use of noise generation and dropout methods.

Most of the existing models for document-level machine translation adopt dual-encoder structures. The representation of the source sentences and the document-level contexts are modeled with two separate encoders. Although these models can make use of the document-level contexts, they do not fully model the interaction between the contexts and the source sentences, and can not directly adapt to the recent pre-training models (e.g., BERT) which encodes multiple sentences with a single encoder. In this work, we propose a simple and effective unified encoder that can outperform the baseline models of dual-encoder models in terms of BLEU and METEOR scores. Moreover, the pre-training models can further boost the performance of our proposed model.

The commonly used framework for unsupervised machine translation builds initial translation models of both translation directions, and then performs iterative back-translation to jointly boost their translation performance. The initialization stage is very important since bad initialization may wrongly squeeze the search space, and too much noise introduced in this stage may hurt the final performance. In this paper, we propose a novel retrieval and rewriting based method to better initialize unsupervised translation models. We first retrieve semantically comparable sentences from monolingual corpora of two languages and then rewrite the target side to minimize the semantic gap between the source and retrieved targets with a designed rewriting model. The rewritten sentence pairs are used to initialize SMT models which are used to generate pseudo data for two NMT models, followed by the iterative back-translation. Experiments show that our method can build better initial unsupervised translation models and improve the final translation performance by over 4 BLEU scores. Our code is released at https://github.com/Imagist-Shuo/RRforUNMT.git.

Neural machine translation systems tend to fail on less decent inputs despite its significant efficacy, which may significantly harm the credibility of these systems—fathoming how and when neural-based systems fail in such cases is critical for industrial maintenance. Instead of collecting and analyzing bad cases using limited handcrafted error features, here we investigate this issue by generating adversarial examples via a new paradigm based on reinforcement learning. Our paradigm could expose pitfalls for a given performance metric, e.g., BLEU, and could target any given neural machine translation architecture. We conduct experiments of adversarial attacks on two mainstream neural machine translation architectures, RNN-search, and Transformer. The results show that our method efficiently produces stable attacks with meaning-preserving adversarial examples. We also present a qualitative and quantitative analysis for the preference pattern of the attack, demonstrating its capability of pitfall exposure.