EMNLP 2021 Papers

One of the most challenging aspects of current single-document news summarization is that the summary often contains ‘extrinsic hallucinations’, i.e., facts that are not present in the source document, which are often derived via world knowledge. This causes summarisation systems to act more like open-ended language models tending to hallucinate facts that are erroneous. In this paper, we mitigate this problem with the help of multiple supplementary resource documents assisting the task. We present a new dataset MiraNews and benchmark existing summarisation models. In contrast to multi-document summarization, which addresses multiple events from several source documents, we still aim at generating a summary for a single document. We show via data analysis that it’s not only the models which are to blame: more than 27% of facts mentioned in the gold summaries of MiraNews are better grounded on assisting documents than in the main source articles. An error analysis of generated summaries from pretrained models fine-tuned on MIRANEWS reveals that this has an even bigger effects on models: assisted summarisation reduces 55% of hallucinations when compared to single-document summarisation models trained on the main article only.

Sequence labeling aims to predict a fine-grained sequence of labels for the text. However, such formulation hinders the effectiveness of supervised methods due to the lack of token-level annotated data. This is exacerbated when we meet a diverse range of languages. In this work, we explore multilingual sequence labeling with minimal supervision using a single unified model for multiple languages. Specifically, we propose a Meta Teacher-Student (MetaTS) Network, a novel meta learning method to alleviate data scarcity by leveraging large multilingual unlabeled data. Prior teacher-student frameworks of self-training rely on rigid teaching strategies, which may hardly produce high-quality pseudo-labels for consecutive and interdependent tokens. On the contrary, MetaTS allows the teacher to dynamically adapt its pseudo-annotation strategies by the student’s feedback on the generated pseudo-labeled data of each language and thus mitigate error propagation from noisy pseudo-labels. Extensive experiments on both public and real-world multilingual sequence labeling datasets empirically demonstrate the effectiveness of MetaTS.

Large-scale multi-label text classification (LMTC) tasks often face long-tailed label distributions, where many labels have few or even no training instances. Although current methods can exploit prior knowledge to handle these few/zero-shot labels, they neglect the meta-knowledge contained in the dataset that can guide models to learn with few samples. In this paper, for the first time, this problem is addressed from a meta-learning perspective. However, the simple extension of meta-learning approaches to multi-label classification is sub-optimal for LMTC tasks due to long-tailed label distribution and coexisting of few- and zero-shot scenarios. We propose a meta-learning approach named META-LMTC. Specifically, it constructs more faithful and more diverse tasks according to well-designed sampling strategies and directly incorporates the objective of adapting to new low-resource tasks into the meta-learning phase. Extensive experiments show that META-LMTC achieves state-of-the-art performance against strong baselines and can still enhance powerful BERTlike models.

With the wide availability of Pre-trained Language Models (PLMs), multi-task fine-tuning across domains has been extensively applied. For tasks related to distant domains with different class label sets, PLMs may memorize non-transferable knowledge for the target domain and suffer from negative transfer. Inspired by meta-learning, we propose the Meta Distant Transfer Learning (Meta-DTL) framework to learn the cross-task knowledge for PLM-based methods. Meta-DTL first employs task representation learning to mine implicit relations among multiple tasks and classes. Based on the results, it trains a PLM-based meta-learner to capture the transferable knowledge across tasks. The weighted maximum entropy regularizers are proposed to make meta-learner more task-agnostic and unbiased. Finally, the meta-learner can be fine-tuned to fit each task with better parameter initialization. We evaluate Meta-DTL using both BERT and ALBERT on seven public datasets. Experiment results confirm the superiority of Meta-DTL as it consistently outperforms strong baselines. We find that Meta-DTL is highly effective when very few data is available for the target task.

Measuring event salience is essential in the understanding of stories. This paper takes a recent unsupervised method for salience detection derived from Barthes Cardinal Functions and theories of surprise and applies it to longer narrative forms. We improve the standard transformer language model by incorporating an external knowledgebase (derived from Retrieval Augmented Generation) and adding a memory mechanism to enhance performance on longer works. We use a novel approach to derive salience annotation using chapter-aligned summaries from the Shmoop corpus for classic literary works. Our evaluation against this data demonstrates that our salience detection model improves performance over and above a non-knowledgebase and memory augmented language model, both of which are crucial to this improvement.

We present MeetDot, a videoconferencing system with live translation captions overlaid on screen. The system aims to facilitate conversation between people who speak different languages, thereby reducing communication barriers between multilingual participants. Currently, our system supports speech and captions in 4 languages and combines automatic speech recognition (ASR) and machine translation (MT) in a cascade. We use the re-translation strategy to translate the streamed speech, resulting in caption flicker. Additionally, our system has very strict latency requirements to have acceptable call quality. We implement several features to enhance user experience and reduce their cognitive load, such as smooth scrolling captions and reducing caption flicker. The modular architecture allows us to integrate different ASR and MT services in our backend. Our system provides an integrated evaluation suite to optimize key intrinsic evaluation metrics such as accuracy, latency and erasure. Finally, we present an innovative cross-lingual word-guessing game as an extrinsic evaluation metric to measure end-to-end system performance. We plan to make our system open-source for research purposes.

Certainty and uncertainty are fundamental to science communication. Hedges have widely been used as proxies for uncertainty. However, certainty is a complex construct, with authors expressing not only the degree but the type and aspects of uncertainty in order to give the reader a certain impression of what is known. Here, we introduce a new study of certainty that models both the level and the aspects of certainty in scientific findings. Using a new dataset of 2167 annotated scientific findings, we demonstrate that hedges alone account for only a partial explanation of certainty. We show that both the overall certainty and individual aspects can be predicted with pre-trained language models, providing a more complete picture of the author’s intended communication. Downstream analyses on 431K scientific findings from news and scientific abstracts demonstrate that modeling sentence-level and aspect-level certainty is meaningful for areas like science communication. Both the model and datasets used in this paper are released at https://blablablab.si.umich.edu/projects/certainty/.

In interpretable NLP, we require faithful rationales that reflect the model’s decision-making process for an explained instance. While prior work focuses on extractive rationales (a subset of the input words), we investigate their less-studied counterpart: free-text natural language rationales. We demonstrate that *pipelines*, models for faithful rationalization on information-extraction style tasks, do not work as well on “reasoning” tasks requiring free-text rationales. We turn to models that *jointly* predict and rationalize, a class of widely used high-performance models for free-text rationalization. We investigate the extent to which the labels and rationales predicted by these models are associated, a necessary property of faithful explanation. Via two tests, *robustness equivalence* and *feature importance agreement*, we find that state-of-the-art T5-based joint models exhibit desirable properties for explaining commonsense question-answering and natural language inference, indicating their potential for producing faithful free-text rationales.

Much of natural language processing is focused on leveraging large capacity language models, typically trained over single messages with a task of predicting one or more tokens. However, modeling human language at higher-levels of context (i.e., sequences of messages) is under-explored. In stance detection and other social media tasks where the goal is to predict an attribute of a message, we have contextual data that is loosely semantically connected by authorship. Here, we introduce Message-Level Transformer (MeLT) – a hierarchical message-encoder pre-trained over Twitter and applied to the task of stance prediction. We focus on stance prediction as a task benefiting from knowing the context of the message (i.e., the sequence of previous messages). The model is trained using a variant of masked-language modeling; where instead of predicting tokens, it seeks to generate an entire masked (aggregated) message vector via reconstruction loss. We find that applying this pre-trained masked message-level transformer to the downstream task of stance detection achieves F1 performance of 67%.

Open Information Extraction (OpenIE) aims to discover textual facts from a given sentence. In essence, the facts contained in plain text are unordered. However, the popular OpenIE systems usually output facts sequentially in the way of predicting the next fact conditioned on the previous decoded ones, which enforce an unnecessary order on the facts and involve the error accumulation between autoregressive steps. To break this bottleneck, we propose MacroIE, a novel non-autoregressive framework for OpenIE. MacroIE firstly constructs a fact graph based on the table filling scheme, in which each node denotes a fact element, and an edge links two nodes that belong to the same fact. Then OpenIE can be reformulated as a non-parametric process of finding maximal cliques from the graph. It directly outputs the final set of facts in one go, thus getting rid of the burden of predicting fact order, as well as the error propagation between facts. Experiments conducted on two benchmark datasets show that our proposed model significantly outperforms current state-of-the-art methods, beats the previous systems by as much as 5.7 absolute gain in F1 score.

There is an increasing interest in the use of mathematical word problem (MWP) generation in educational assessment. Different from standard natural question generation, MWP generation needs to maintain the underlying mathematical operations between quantities and variables, while at the same time ensuring the relevance between the output and the given topic. To address above problem, we develop an end-to-end neural model to generate diverse MWPs in real-world scenarios from commonsense knowledge graph and equations. The proposed model (1) learns both representations from edge-enhanced Levi graphs of symbolic equations and commonsense knowledge; (2) automatically fuses equation and commonsense knowledge information via a self-planning module when generating the MWPs. Experiments on an educational gold-standard set and a large-scale generated MWP set show that our approach is superior on the MWP generation task, and it outperforms the SOTA models in terms of both automatic evaluation metrics, i.e., BLEU-4, ROUGE-L, Self-BLEU, and human evaluation metrics, i.e., equation relevance, topic relevance, and language coherence. To encourage reproducible results, we make our code and MWP dataset public available at https://github.com/tal-ai/MaKE_EMNLP2021.

We study the problem of generating arithmetic math word problems (MWPs) given a math equation that specifies the mathematical computation and a context that specifies the problem scenario. Existing approaches are prone to generating MWPs that are either mathematically invalid or have unsatisfactory language quality. They also either ignore the context or require manual specification of a problem template, which compromises the diversity of the generated MWPs. In this paper, we develop a novel MWP generation approach that leverages i) pre-trained language models and a context keyword selection model to improve the language quality of generated MWPs and ii) an equation consistency constraint for math equations to improve the mathematical validity of the generated MWPs. Extensive quantitative and qualitative experiments on three real-world MWP datasets demonstrate the superior performance of our approach compared to various baselines.

Product quantization (PQ) is a widely used technique for ad-hoc retrieval. Recent studies propose supervised PQ, where the embedding and quantization models can be jointly trained with supervised learning. However, there is a lack of appropriate formulation of the joint training objective; thus, the improvements over previous non-supervised baselines are limited in reality. In this work, we propose the Matching-oriented Product Quantization (MoPQ), where a novel objective Multinoulli Contrastive Loss (MCL) is formulated. With the minimization of MCL, we are able to maximize the matching probability of query and ground-truth key, which contributes to the optimal retrieval accuracy. Given that the exact computation of MCL is intractable due to the demand of vast contrastive samples, we further propose the Differentiable Cross-device Sampling (DCS), which significantly augments the contrastive samples for precise approximation of MCL. We conduct extensive experimental studies on four real-world datasets, whose results verify the effectiveness of MoPQ. The code is available at https://github.com/microsoft/MoPQ.

Current research in automatic summarisation is unapologetically anglo-centered–a persistent state-of-affairs, which also predates neural net approaches. High-quality automatic summarisation datasets are notoriously expensive to create, posing a challenge for any language. However, with digitalisation, archiving, and social media advertising of newswire articles, recent work has shown how, with careful methodology application, large-scale datasets can now be simply gathered instead of written. In this paper, we present a large-scale multilingual summarisation dataset containing articles in 92 languages, spread across 28.8 million articles, in more than 35 writing scripts. This is both the largest, most inclusive, existing automatic summarisation dataset, as well as one of the largest, most inclusive, ever published datasets for any NLP task. We present the first investigation on the efficacy of resource building from news platforms in the low-resource language setting. Finally, we provide some first insight on how low-resource language settings impact state-of-the-art automatic summarisation system performance.

A possible explanation for the impressive performance of masked language model (MLM) pre-training is that such models have learned to represent the syntactic structures prevalent in classical NLP pipelines. In this paper, we propose a different explanation: MLMs succeed on downstream tasks almost entirely due to their ability to model higher-order word co-occurrence statistics. To demonstrate this, we pre-train MLMs on sentences with randomly shuffled word order, and show that these models still achieve high accuracy after fine-tuning on many downstream tasks—including tasks specifically designed to be challenging for models that ignore word order. Our models perform surprisingly well according to some parametric syntactic probes, indicating possible deficiencies in how we test representations for syntactic information. Overall, our results show that purely distributional information largely explains the success of pre-training, and underscore the importance of curating challenging evaluation datasets that require deeper linguistic knowledge.

While solving math word problems automatically has received considerable attention in the NLP community, few works have addressed probability word problems specifically. In this paper, we employ and analyse various neural models for answering such word problems. In a two-step approach, the problem text is first mapped to a formal representation in a declarative language using a sequence-to-sequence model, and then the resulting representation is executed using a probabilistic programming system to provide the answer. Our best performing model incorporates general-domain contextualised word representations that were finetuned using transfer learning on another in-domain dataset. We also apply end-to-end models to this task, which bring out the importance of the two-step approach in obtaining correct solutions to probability problems.

Mapping natural language instructions to programs that computers can process is a fundamental challenge. Existing approaches focus on likelihood-based training or using reinforcement learning to fine-tune models based on a single reward. In this paper, we pose program generation from language as Inverse Reinforcement Learning. We introduce several interpretable reward components and jointly learn (1) a reward function that linearly combines them, and (2) a policy for program generation. Fine-tuning with our approach achieves significantly better performance than competitive methods using Reinforcement Learning (RL). On the VirtualHome framework, we get improvements of up to 9.0% on the Longest Common Subsequence metric and 14.7% on recall-based metrics over previous work on this framework (Puig et al., 2018). The approach is data-efficient, showing larger gains in performance in the low-data regime. Generated programs are also preferred by human evaluators over an RL-based approach, and rated higher on relevance, completeness, and human-likeness.

Neural relation extraction models have shown promising results in recent years; however, the model performance drops dramatically given only a few training samples. Recent works try leveraging the advance in few-shot learning to solve the low resource problem, where they train label-agnostic models to directly compare the semantic similarities among context sentences in the embedding space. However, the label-aware information, i.e., the relation label that contains the semantic knowledge of the relation itself, is often neglected for prediction. In this work, we propose a framework considering both label-agnostic and label-aware semantic mapping information for low resource relation extraction. We show that incorporating the above two types of mapping information in both pretraining and fine-tuning can significantly improve the model performance on low-resource relation extraction tasks.

For interpreting the behavior of a probabilistic model, it is useful to measure a model’s calibration—the extent to which it produces reliable confidence scores. We address the open problem of calibration for tagging models with sparse tagsets, and recommend strategies to measure and reduce calibration error (CE) in such models. We show that several post-hoc recalibration techniques all reduce calibration error across the marginal distribution for two existing sequence taggers. Moreover, we propose tag frequency grouping (TFG) as a way to measure calibration error in different frequency bands. Further, recalibrating each group separately promotes a more equitable reduction of calibration error across the tag frequency spectrum.