EMNLP 2021 论文列表

Asking questions about a situation is an inherent step towards understanding it. To this end, we introduce the task of role question generation, which, given a predicate mention and a passage, requires producing a set of questions asking about all possible semantic roles of the predicate. We develop a two-stage model for this task, which first produces a context-independent question prototype for each role and then revises it to be contextually appropriate for the passage. Unlike most existing approaches to question generation, our approach does not require conditioning on existing answers in the text. Instead, we condition on the type of information to inquire about, regardless of whether the answer appears explicitly in the text, could be inferred from it, or should be sought elsewhere. Our evaluation demonstrates that we generate diverse and well-formed questions for a large, broad-coverage ontology of predicates and roles.

The impressive capabilities of recent generative models to create texts that are challenging to distinguish from the human-written ones can be misused for generating fake news, product reviews, and even abusive content. Despite the prominent performance of existing methods for artificial text detection, they still lack interpretability and robustness towards unseen models. To this end, we propose three novel types of interpretable topological features for this task based on Topological Data Analysis (TDA) which is currently understudied in the field of NLP. We empirically show that the features derived from the BERT model outperform count- and neural-based baselines up to 10% on three common datasets, and tend to be the most robust towards unseen GPT-style generation models as opposed to existing methods. The probing analysis of the features reveals their sensitivity to the surface and syntactic properties. The results demonstrate that TDA is a promising line with respect to NLP tasks, specifically the ones that incorporate surface and structural information.

Evidence association in criminal cases is dividing a set of judicial evidence into several non-overlapping subsets, improving the interpretability and legality of conviction. Observably, evidence divided into the same subset usually supports the same claim. Therefore, we propose an argumentation-driven supervised learning method to calculate the distance between evidence pairs for the following evidence association step in this paper. Experimental results on a real-world dataset demonstrate the effectiveness of our method.

Argument pair extraction (APE) aims to extract interactive argument pairs from two passages of a discussion. Previous work studied this task in the context of peer review and rebuttal, and decomposed it into a sequence labeling task and a sentence relation classification task. However, despite the promising performance, such an approach obtains the argument pairs implicitly by the two decomposed tasks, lacking explicitly modeling of the argument-level interactions between argument pairs. In this paper, we tackle the APE task by a mutual guidance framework, which could utilize the information of an argument in one passage to guide the identification of arguments that can form pairs with it in another passage. In this manner, two passages can mutually guide each other in the process of APE. Furthermore, we propose an inter-sentence relation graph to effectively model the inter-relations between two sentences and thus facilitates the extraction of argument pairs. Our proposed method can better represent the holistic argument-level semantics and thus explicitly capture the complex correlations between argument pairs. Experimental results show that our approach significantly outperforms the current state-of-the-art model.

Many recent works have demonstrated that unsupervised sentence representations of neural networks encode syntactic information by observing that neural language models are able to predict the agreement between a verb and its subject. We take a critical look at this line of research by showing that it is possible to achieve high accuracy on this agreement task with simple surface heuristics, indicating a possible flaw in our assessment of neural networks’ syntactic ability. Our fine-grained analyses of results on the long-range French object-verb agreement show that contrary to LSTMs, Transformers are able to capture a non-trivial amount of grammatical structure.

Internet search affects people’s cognition of the world, so mitigating biases in search results and learning fair models is imperative for social good. We study a unique gender bias in image search in this work: the search images are often gender-imbalanced for gender-neutral natural language queries. We diagnose two typical image search models, the specialized model trained on in-domain datasets and the generalized representation model pre-trained on massive image and text data across the internet. Both models suffer from severe gender bias. Therefore, we introduce two novel debiasing approaches: an in-processing fair sampling method to address the gender imbalance issue for training models, and a post-processing feature clipping method base on mutual information to debias multimodal representations of pre-trained models. Extensive experiments on MS-COCO and Flickr30K benchmarks show that our methods significantly reduce the gender bias in image search models.

With the continuous upgrading of the summarization systems driven by deep neural networks, researchers have higher requirements on the quality of the generated summaries, which should be not only fluent and informative but also factually correct. As a result, the field of factual evaluation has developed rapidly recently. Despite its initial progress in evaluating generated summaries, the meta-evaluation methodologies of factuality metrics are limited in their opacity, leading to the insufficient understanding of factuality metrics’ relative advantages and their applicability. In this paper, we present an adversarial meta-evaluation methodology that allows us to (i) diagnose the fine-grained strengths and weaknesses of 6 existing top-performing metrics over 24 diagnostic test datasets, (ii) search for directions for further improvement by data augmentation. Our observations from this work motivate us to propose several calls for future research. We make all codes, diagnostic test datasets, trained factuality models available: https://github.com/zide05/AdvFact.

Pre-training Transformer-based models such as BERT and ELECTRA on a collection of Arabic corpora, demonstrated by both AraBERT and AraELECTRA, shows an impressive result on downstream tasks. However, pre-training Transformer-based language models is computationally expensive, especially for large-scale models. Recently, Funnel Transformer has addressed the sequential redundancy inside Transformer architecture by compressing the sequence of hidden states, leading to a significant reduction in the pre-training cost. This paper empirically studies the performance and efficiency of building an Arabic language model with Funnel Transformer and ELECTRA objective. We find that our model achieves state-of-the-art results on several Arabic downstream tasks despite using less computational resources compared to other BERT-based models.

We develop a unified system to answer directly from text open-domain questions that may require a varying number of retrieval steps. We employ a single multi-task transformer model to perform all the necessary subtasks—retrieving supporting facts, reranking them, and predicting the answer from all retrieved documents—in an iterative fashion. We avoid crucial assumptions of previous work that do not transfer well to real-world settings, including exploiting knowledge of the fixed number of retrieval steps required to answer each question or using structured metadata like knowledge bases or web links that have limited availability. Instead, we design a system that can answer open-domain questions on any text collection without prior knowledge of reasoning complexity. To emulate this setting, we construct a new benchmark, called BeerQA, by combining existing one- and two-step datasets with a new collection of 530 questions that require three Wikipedia pages to answer, unifying Wikipedia corpora versions in the process. We show that our model demonstrates competitive performance on both existing benchmarks and this new benchmark. We make the new benchmark available at https://beerqa.github.io/.

Word embeddings are powerful representations that form the foundation of many natural language processing architectures, both in English and in other languages. To gain further insight into word embeddings, we explore their stability (e.g., overlap between the nearest neighbors of a word in different embedding spaces) in diverse languages. We discuss linguistic properties that are related to stability, drawing out insights about correlations with affixing, language gender systems, and other features. This has implications for embedding use, particularly in research that uses them to study language trends.

We analyze language change over time in a collaborative, goal-oriented instructional task, where utility-maximizing participants form conventions and increase their expertise. Prior work studied such scenarios mostly in the context of reference games, and consistently found that language complexity is reduced along multiple dimensions, such as utterance length, as conventions are formed. In contrast, we find that, given the ability to increase instruction utility, instructors increase language complexity along these previously studied dimensions to better collaborate with increasingly skilled instruction followers.

Morality plays an important role in social well-being, but people’s moral perception is not stable and changes over time. Recent advances in natural language processing have shown that text is an effective medium for informing moral change, but no attempt has been made to quantify the origins of these changes. We present a novel unsupervised framework for tracing textual sources of moral change toward entities through time. We characterize moral change with probabilistic topical distributions and infer the source text that exerts prominent influence on the moral time course. We evaluate our framework on a diverse set of data ranging from social media to news articles. We show that our framework not only captures fine-grained human moral judgments, but also identifies coherent source topics of moral change triggered by historical events. We apply our methodology to analyze the news in the COVID-19 pandemic and demonstrate its utility in identifying sources of moral change in high-impact and real-time social events.

Online conversations include more than just text. Increasingly, image-based responses such as memes and animated gifs serve as culturally recognized and often humorous responses in conversation. However, while NLP has broadened to multimodal models, conversational dialog systems have largely focused only on generating text replies. Here, we introduce a new dataset of 1.56M text-gif conversation turns and introduce a new multimodal conversational model Pepe the King Prawn for selecting gif-based replies. We demonstrate that our model produces relevant and high-quality gif responses and, in a large randomized control trial of multiple models replying to real users, we show that our model replies with gifs that are significantly better received by the community.

The availability of parallel sentence simplification (SS) is scarce for neural SS modelings. We propose an unsupervised method to build SS corpora from large-scale bilingual translation corpora, alleviating the need for SS supervised corpora. Our method is motivated by the following two findings: neural machine translation model usually tends to generate more high-frequency tokens and the difference of text complexity levels exists between the source and target language of a translation corpus. By taking the pair of the source sentences of translation corpus and the translations of their references in a bridge language, we can construct large-scale pseudo parallel SS data. Then, we keep these sentence pairs with a higher complexity difference as SS sentence pairs. The building SS corpora with an unsupervised approach can satisfy the expectations that the aligned sentences preserve the same meanings and have difference in text complexity levels. Experimental results show that SS methods trained by our corpora achieve the state-of-the-art results and significantly outperform the results on English benchmark WikiLarge.

Aspect detection is a fundamental task in opinion mining. Previous works use seed words either as priors of topic models, as anchors to guide the learning of aspects, or as features of aspect classifiers. This paper presents a novel weakly-supervised method to exploit seed words for aspect detection based on an encoder architecture. The encoder maps segments and aspects into a low-dimensional embedding space. The goal is approximating similarity between segments and aspects in the embedding space and their ground-truth similarity generated from seed words. An objective function is proposed to capture the uncertainty of ground-truth similarity. Our method outperforms previous works on several benchmarks in various domains.

Aspect-based sentiment analysis (ABSA) mainly involves three subtasks: aspect term extraction, opinion term extraction, and aspect-level sentiment classification, which are typically handled in a separate or joint manner. However, previous approaches do not well exploit the interactive relations among three subtasks and do not pertinently leverage the easily available document-level labeled domain/sentiment knowledge, which restricts their performances. To address these issues, we propose a novel Iterative Multi-Knowledge Transfer Network (IMKTN) for end-to-end ABSA. For one thing, through the interactive correlations between the ABSA subtasks, our IMKTN transfers the task-specific knowledge from any two of the three subtasks to another one at the token level by utilizing a well-designed routing algorithm, that is, any two of the three subtasks will help the third one. For another, our IMKTN pertinently transfers the document-level knowledge, i.e., domain-specific and sentiment-related knowledge, to the aspect-level subtasks to further enhance the corresponding performance. Experimental results on three benchmark datasets demonstrate the effectiveness and superiority of our approach.

Linguistic typology generally divides synthetic languages into groups based on their morphological fusion. However, this measure has long been thought to be best considered a matter of degree. We present an information-theoretic measure, called informational fusion, to quantify the degree of fusion of a given set of morphological features in a surface form, which naturally provides such a graded scale. Informational fusion is able to encapsulate not only concatenative, but also nonconcatenative morphological systems (e.g. Arabic), abstracting away from any notions of morpheme segmentation. We then show, on a sample of twenty-one languages, that our measure recapitulates the usual linguistic classifications for concatenative systems, and provides new measures for nonconcatenative ones. We also evaluate the long-standing hypotheses that more frequent forms are more fusional, and that paradigm size anticorrelates with degree of fusion. We do not find evidence for the idea that languages have characteristic levels of fusion; rather, the degree of fusion varies across part-of-speech within languages.

Dialogue summarization helps readers capture salient information from long conversations in meetings, interviews, and TV series. However, real-world dialogues pose a great challenge to current summarization models, as the dialogue length typically exceeds the input limits imposed by recent transformer-based pre-trained models, and the interactive nature of dialogues makes relevant information more context-dependent and sparsely distributed than news articles. In this work, we perform a comprehensive study on long dialogue summarization by investigating three strategies to deal with the lengthy input problem and locate relevant information: (1) extended transformer models such as Longformer, (2) retrieve-then-summarize pipeline models with several dialogue utterance retrieval methods, and (3) hierarchical dialogue encoding models such as HMNet. Our experimental results on three long dialogue datasets (QMSum, MediaSum, SummScreen) show that the retrieve-then-summarize pipeline models yield the best performance. We also demonstrate that the summary quality can be further improved with a stronger retrieval model and pretraining on proper external summarization datasets.

Intent classification (IC) and slot filling (SF) are critical building blocks in task-oriented dialogue systems. These two tasks are closely-related and can flourish each other. Since only a few utterances can be utilized for identifying fast-emerging new intents and slots, data scarcity issue often occurs when implementing IC and SF. However, few IC/SF models perform well when the number of training samples per class is quite small. In this paper, we propose a novel explicit-joint and supervised-contrastive learning framework for few-shot intent classification and slot filling. Its highlights are as follows. (i) The model extracts intent and slot representations via bidirectional interactions, and extends prototypical network to achieve explicit-joint learning, which guarantees that IC and SF tasks can mutually reinforce each other. (ii) The model integrates with supervised contrastive learning, which ensures that samples from same class are pulled together and samples from different classes are pushed apart. In addition, the model follows a not common but practical way to construct the episode, which gets rid of the traditional setting with fixed way and shot, and allows for unbalanced datasets. Extensive experiments on three public datasets show that our model can achieve promising performance.