EMNLP 2022 论文列表

Text data can pose a risk of harm. However, the risks are not fully understood, and how to handle, present, and discuss harmful text in a safe way remains an unresolved issue in the NLP community. We provide an analytical framework categorising harms on three axes: (1) the harm type (e.g., misinformation, hate speech or racial stereotypes); (2) whether a harm is sought as a feature of the research design if explicitly studying harmful content (e.g., training a hate speech classifier), versus unsought if harmful content is encountered when working on unrelated problems (e.g., language generation or part-of-speech tagging); and (3) who it affects, from people (mis)represented in the data to those handling the data and those publishing on the data. We provide advice for practitioners, with concrete steps for mitigating harm in research and in publication. To assist implementation we introduce HarmCheck – a documentation standard for handling and presenting harmful text in research.

Hierarchical text classification (HTC) is a challenging subtask of multi-label classification due to its complex label hierarchy.Recently, the pretrained language models (PLM)have been widely adopted in HTC through a fine-tuning paradigm. However, in this paradigm, there exists a huge gap between the classification tasks with sophisticated label hierarchy and the masked language model (MLM) pretraining tasks of PLMs and thus the potential of PLMs cannot be fully tapped.To bridge the gap, in this paper, we propose HPT, a Hierarchy-aware Prompt Tuning method to handle HTC from a multi-label MLM perspective.Specifically, we construct a dynamic virtual template and label words that take the form of soft prompts to fuse the label hierarchy knowledge and introduce a zero-bounded multi-label cross-entropy loss to harmonize the objectives of HTC and MLM.Extensive experiments show HPT achieves state-of-the-art performances on 3 popular HTC datasets and is adept at handling the imbalance and low resource situations. Our code is available at https://github.com/wzh9969/HPT.

Extractive summarization for long documents is challenging due to the extended structured input context. The long-distance sentence dependency hinders cross-sentence relations modeling, the critical step of extractive summarization. This paper proposes HEGEL, a hypergraph neural network for long document summarization by capturing high-order cross-sentence relations. HEGEL updates and learns effective sentence representations with hypergraph transformer layers and fuses different types of sentence dependencies, including latent topics, keywords coreference, and section structure. We validate HEGEL by conducting extensive experiments on two benchmark datasets, and experimental results demonstrate the effectiveness and efficiency of HEGEL.

Event detection has been suffering from constantly emerging event types with lack of sufficient data. Existing works formulate the new problem as few-shot event detection (FSED), and employ two-stage or unified models based on meta-learning to address the problem. However, these methods fall far short of expectations due to: (i) insufficient learning of discriminative representations in low-resource scenarios, and (ii) representation overlap between triggers and non-triggers. To resolve the above issues, in this paper, we propose a novel Hybrid Contrastive Learning method with a Task-Adaptive Threshold (abbreviated as HCL-TAT), which enables discriminative representation learning with a two-view contrastive loss (support-support and prototype-query), and devises an easily-adapted threshold to alleviate misidentification of triggers. Extensive experiments on the benchmark dataset FewEvent demonstrate the superiority of our method to achieve better results compared to the state-of-the-arts. All the data and codes will be available to facilitate future research.

The successful completion of the hate speech detection task hinges upon the availability of rich and variable labeled data, which is hard to obtain. In this work, we present a new approach for data augmentation that uses as input real unlabelled data, which is carefully selected from online platforms where invited hate speech is abundant. We show that by harvesting and processing this data (in an automatic manner), one can augment existing manually-labeled datasets to improve the classification performance of hate speech classification models. We observed an improvement in F1-score ranging from 2.7% and up to 9.5%, depending on the task (in- or cross-domain) and the model used.

The phenomenon of zero pronoun (ZP) has attracted increasing interest in the machine translation (MT) community due to its importance and difficulty. However, previous studies generally evaluate the quality of translating ZPs with BLEU scores on MT testsets, which is not expressive or sensitive enough for accurate assessment. To bridge the data and evaluation gaps, we propose a benchmark testset for target evaluation on Chinese-English ZP translation. The human-annotated testset covers five challenging genres, which reveal different characteristics of ZPs for comprehensive evaluation. We systematically revisit eight advanced models on ZP translation and identify current challenges for future exploration. We release data, code, models and annotation guidelines, which we hope can significantly promote research in this field (https://github.com/longyuewangdcu/mZPRT).

Automated storytelling has long captured the attention of researchers for the ubiquity of narratives in everyday life. However, it is challenging to maintain coherence and stay on-topictoward a specific ending when generating narratives with neural language models. In this paper, we introduce Story generation with ReaderModels (StoRM), a framework in which areader model is used to reason about the storyshould progress. A reader model infers whata human reader believes about the concepts,entities, and relations about the fictional storyworld. We show how an explicit reader modelrepresented as a knowledge graph affords the storycoherence and provides controllability in theform of achieving a given story world stategoal. Experiments show that our model produces significantly more coherent and on-topicstories, outperforming baselines in dimensionsincluding plot plausibility and staying on topic

Machine Translation task has made great progress with the help of auto-regressive decoding paradigm and Transformer architecture. In this paradigm, though the encoder can obtain global source representations, the decoder can only use translation history to determine the current word. Previous promising works attempted to address this issue by applying a draft or a fixed-length semantic embedding as target-side global information. However, these methods either degrade model efficiency or show limitations in expressing semantics. Motivated by Functional Equivalence Theory, we extract several semantic kernels from a source sentence, each of which can express one semantic segment of the original sentence. Together, these semantic kernels can capture global semantic information, and we project them into target embedding space to guide target sentence generation. We further force our model to use semantic kernels at each decoding step through an adaptive mask algorithm. Empirical studies on various machine translation benchmarks show that our approach gains approximately an improvement of 1 BLEU score on most benchmarks over the Transformer baseline and about 1.7 times faster than previous works on average at inference time.

Entity Alignment (EA) aims to find equivalent entities between two Knowledge Graphs (KGs). While numerous neural EA models have been devised, they are mainly learned using labelled data only. In this work, we argue that different entities within one KG should have compatible counterparts in the other KG due to the potential dependencies among the entities. Making compatible predictions thus should be one of the goals of training an EA model along with fitting the labelled data: this aspect however is neglected in current methods. To power neural EA models with compatibility, we devise a training framework by addressing three problems: (1) how to measure the compatibility of an EA model; (2) how to inject the property of being compatible into an EA model; (3) how to optimise parameters of the compatibility model. Extensive experiments on widely-used datasets demonstrate the advantages of integrating compatibility within EA models. In fact, state-of-the-art neural EA models trained within our framework using just 5% of the labelled data can achieve comparable effectiveness with supervised training using 20% of the labelled data.

Abstractive dialogue summarization has recently been receiving more attention. We propose a coarse-to-fine model for generating abstractive dialogue summaries, and introduce a fact-aware reinforcement learning (RL) objective that improves the fact consistency between the dialogue and the generated summary. Initially, the model generates the predicate-argument spans of the dialogue, and then generates the final summary through a fact-aware RL objective. Extensive experiments and analysis on two benchmark datasets demonstrate that our proposed method effectively improves the quality of the generated summary, especially in coherence and consistency.

Recent joint multiple intent detection and slot filling models employ label embeddings to achieve the semantics-label interactions.However, they treat all labels and label embeddings as uncorrelated individuals, ignoring the dependencies among them. Besides, they conduct the decoding for the two tasks independently, without leveraging the correlations between them.Therefore, in this paper, we first construct a Heterogeneous Label Graph (HLG) containing two kinds of topologies: (1) statistical dependencies based on labels’ co-occurrence patterns and hierarchies in slot labels; (2) rich relations among the label nodes.Then we propose a novel model termed ReLa-Net.It can capture beneficial correlations among the labels from HLG.The label correlations are leveraged to enhance semantic-label interactions. Moreover, we also propose the label-aware inter-dependent decoding mechanism to further exploit the label correlations for decoding. Experiment results show that our ReLa-Net significantly outperforms previous models.Remarkably, ReLa-Net surpasses the previous best model by over 20% in terms of overall accuracy on MixATIS dataset.

Large pre-trained language models have recently enabled open-ended generation frameworks (e.g., prompt-to-text NLG) to tackle a variety of tasks going beyond the traditional data-to-text generation. While this framework is more general, it is under-specified and often leads to a lack of controllability restricting their real-world usage. We propose a new grounded keys-to-text generation task: the task is to generate a factual description about an entity given a set of guiding keys, and grounding passages. To address this task, we introduce a new dataset, called EntDeGen. Inspired by recent QA-based evaluation measures, we propose an automatic metric, MAFE, for factual correctness of generated descriptions. Our EntDescriptor model is equipped with strong rankers to fetch helpful passages and generate entity descriptions. Experimental result shows a good correlation (60.14) between our proposed metric and human judgments of factuality. Our rankers significantly improved the factual correctness of generated descriptions (15.95% and 34.51% relative gains in recall and precision). Finally, our ablation study highlights the benefit of combining keys and groundings.

Despite recent explosion of interests in in-context learning, the underlying mechanism and the precise impact of the quality of demonstrations remain elusive.Intuitively, ground-truth labels should have as much impact in in-context learning (ICL) as supervised learning, but recent work reported that the input-label correspondence is significantly less important than previously thought.Intrigued by this counter-intuitive observation, we re-examine the importance of ground-truth labels in in-context learning.With the introduction of two novel metrics, namely Label-Correctness Sensitivity and Ground-truth Label Effect Ratio (GLER), we were able to conduct quantifiable analysis on the impact of ground-truth label demonstrations.Through extensive analyses, we find that the correct input-label mappings can have varying impacts on the downstream in-context learning performances, depending on the experimental configuration.Through additional studies, we identify key components, such as the verbosity of prompt templates and the language model size, as the controlling factor to achieve more noise-resilient ICL.

Subject to the huge semantic gap between natural and formal languages, neural semantic parsing is typically bottlenecked by its complexity of dealing with both input semantics and output syntax. Recent works have proposed several forms of supplementary supervision but none is generalized across multiple formal languages. This paper proposes a unified intermediate representation for graph query languages, named GraphQ IR. It has a natural-language-like expression that bridges the semantic gap and formally defined syntax that maintains the graph structure. Therefore, a neural semantic parser can more precisely convert user queries into GraphQ IR, which can be later losslessly compiled into various downstream graph query languages. Extensive experiments on several benchmarks including KQA Pro, Overnight, GrailQA, and MetaQA-Cypher under the standard i.i.d., out-of-distribution, and low-resource settings validate GraphQ IR’s superiority over the previous state-of-the-arts with a maximum 11% accuracy improvement.

Few-shot relation extraction (FSRE) has been a challenging problem since it only has a handful of training instances. Existing models follow a ‘one-for-all’ scheme where one general large model performs all individual N-way-K-shot tasks in FSRE, which prevents the model from achieving the optimal point on each task. In view of this, we propose a model generation framework that consists of one general model for all tasks and many tiny task-specific models for each individual task. The general model generates and passes the universal knowledge to the tiny models which will be further fine-tuned when performing specific tasks. In this way, we decouple the complexity of the entire task space from that of all individual tasks while absorbing the universal knowledge.Extensive experimental results on two public datasets demonstrate that our framework reaches a new state-of-the-art performance for FRSE tasks. Our code is available at: https://github.com/NLPWM-WHU/GM_GEN.

A graph is a suitable data structure to represent the structural information of text. Recently, multi-hop question answering (MHQA) tasks, which require inter-paragraph/sentence linkages, have come to exploit such properties of a graph. Previous approaches to MHQA relied on leveraging the graph information along with the pre-trained language model (PLM) encoders. However, this trend exhibits the following drawbacks: (i) sample inefficiency while training in a low-resource setting; (ii) lack of reusability due to changes in the model structure or input. Our work proposes the Graph-Induced Transformer (GIT) that applies graph-derived attention patterns directly into a PLM, without the need to employ external graph modules. GIT can leverage the useful inductive bias of graphs while retaining the unperturbed Transformer structure and parameters. Our experiments on HotpotQA successfully demonstrate both the sample efficient characteristic of GIT and its capacity to replace the graph modules while preserving model performance.

Part-of-Speech (POS) tagging is an important component of the NLP pipeline, but many low-resource languages lack labeled data for training. An established method for training a POS tagger in such a scenario is to create a labeled training set by transferring from high-resource languages. In this paper, we propose a novel method for transferring labels from multiple high-resource source to low-resource target languages. We formalize POS tag projection as graph-based label propagation. Given translations of a sentence in multiple languages, we create a graph with words as nodes and alignment links as edges by aligning words for all language pairs. We then propagate node labels from source to target using a Graph Neural Network augmented with transformer layers. We show that our propagation creates training sets that allow us to train POS taggers for a diverse set of languages. When combined with enhanced contextualized embeddings, our method achieves a new state-of-the-art for unsupervised POS tagging of low-resource languages.

Temporal Knowledge Graph (TKG) reasoning has attracted increasing attention due to its enormous potential value, and the critical issue is how to model the complex temporal structure information effectively. Recent studies use the method of encoding graph snapshots into hidden vector space and then performing heuristic deductions, which perform well on the task of entity prediction. However, these approaches cannot predict when an event will occur and have the following limitations: 1) there are many facts not related to the query that can confuse the model; 2) there exists information forgetting caused by long-term evolutionary processes. To this end, we propose a Graph Hawkes Transformer (GHT) for both TKG entity prediction and time prediction tasks in the future time. In GHT, there are two variants of Transformer, which capture the instantaneous structural information and temporal evolution information, respectively, and a new relational continuous-time encoding function to facilitate feature evolution with the Hawkes process. Extensive experiments on four public datasets demonstrate its superior performance, especially on long-term evolutionary tasks.

Argumentation is used by people both internally, by evaluating arguments and counterarguments to make sense of a situation and take a decision, and externally, e.g., in a debate, by exchanging arguments to reach an agreement or to promote an individual position. In this context, the assessment of the quality of the arguments is of extreme importance, as it strongly influences the evaluation of the overall argumentation, impacting on the decision making process. The automatic assessment of the quality of natural language arguments is recently attracting interest in the Argument Mining field. However, the issue of automatically assessing the quality of an argumentation largely remains a challenging unsolved task. Our contribution is twofold: first, we present a novel resource of 402 student persuasive essays, where three main quality dimensions (i.e., cogency, rhetoric, and reasonableness) have been annotated, leading to 1908 arguments tagged with quality facets; second, we address this novel task of argumentation quality assessment proposing a novel neural architecture based on graph embeddings, that combines both the textual features of the natural language arguments and the overall argument graph, i.e., considering also the support and attack relations holding among the arguments. Results on the persuasive essays dataset outperform state-of-the-art and standard baselines’ performance.