COLING 2022 Papers

Event extraction (EE) is one of the fundamental tasks for information extraction whose goal is to identify mentions of events and their participants in text. Due to its importance, different methods and datasets have been introduced for EE. However, existing EE datasets are limited to formally written documents such as news articles or scientific papers. As such, the challenges of EE in informal and noisy texts are not adequately studied. In particular, video transcripts constitute an important domain that can benefit tremendously from EE systems (e.g., video retrieval), but has not been studied in EE literature due to the lack of necessary datasets. To address this limitation, we propose the first large-scale EE dataset obtained for transcripts of streamed videos on the video hosting platform Behance to promote future research in this area. In addition, we extensively evaluate existing state-of-the-art EE methods on our new dataset. We demonstrate that such systems cannot achieve adequate performance on the proposed dataset, revealing challenges and opportunities for further research effort.

Event detection, which aims to identify instances of specific event types from pieces of text, is a fundamental task in information extraction. Most existing approaches leverage syntactic knowledge with a set of syntactic relations to enhance event detection. However, a side effect of these syntactic-based approaches is that they may confuse different syntactic relations and tend to introduce redundant or noisy information, which may lead to performance degradation. To this end, we propose a simple yet effective model named DualGAT (Dual Relational Graph Attention Networks), which exploits the complementary nature of syntactic and semantic relations to alleviate the problem. Specifically, we first construct a dual relational graph that both aggregates syntactic and semantic relations to the key nodes in the graph, so that event-relevant information can be comprehensively captured from multiple perspectives (i.e., syntactic and semantic views). We then adopt augmented relational graph attention networks to encode the graph and optimize its attention weights by introducing contextual information, which further improves the performance of event detection. Extensive experiments conducted on the standard ACE2005 benchmark dataset indicate that our method significantly outperforms the state-of-the-art methods and verifies the superiority of DualGAT over existing syntactic-based methods.

This paper studies event causality identification, which aims at predicting the causality relation for a pair of events in a sentence. Regarding event causality identification as a supervised classification task, most existing methods suffer from the problem of insufficient annotated data. In this paper, we propose a new derivative prompt joint learning model for event causality identification, which leverages potential causal knowledge in the pre-trained language model to tackle the data scarcity problem. Specifically, rather than external data or knowledge augmentation, we derive two relevant prompt tasks from event causality identification to enhance the model’s ability to identify explicit and implicit causality. We evaluate our model on two benchmark datasets and the results show that our model has great advantages over previous methods.

Event Causality Identification (ECI), which aims to detect whether a causality relation exists between two given textual events, is an important task for event causality understanding. However, the ECI task ignores crucial event structure and cause-effect causality component information, making it struggle for downstream applications. In this paper, we introduce a novel task, namely Event Causality Extraction (ECE), aiming to extract the cause-effect event causality pairs with their structured event information from plain texts. The ECE task is more challenging since each event can contain multiple event arguments, posing fine-grained correlations between events to decide the cause-effect event pair. Hence, we propose a method with a dual grid tagging scheme to capture the intra- and inter-event argument correlations for ECE. Further, we devise a event type-enhanced model architecture to realize the dual grid tagging scheme. Experiments demonstrate the effectiveness of our method, and extensive analyses point out several future directions for ECE.

Difficulties with social aspects of language are among the hallmarks of autism spectrum disorder (ASD). These communication differences are thought to contribute to the challenges that adults with ASD experience when seeking employment, underscoring the need for interventions that focus on improving areas of weakness in pragmatic and social language. In this paper, we describe a transformer-based framework for identifying linguistic features associated with social aspects of communication using a corpus of conversations between adults with and without ASD and neurotypical conversational partners produced while engaging in collaborative tasks. While our framework yields strong accuracy overall, performance is significantly worse for the language of participants with ASD, suggesting that they use a more diverse set of strategies for some social linguistic functions. These results, while showing promise for the development of automated language analysis tools to support targeted language interventions for ASD, also reveal weaknesses in the ability of large contextualized language models to model neuroatypical language.

Recent studies show that NLP models trained on standard English texts tend to produce biased outcomes against underrepresented English varieties. In this work, we conduct a pioneering study of the English variety use of African American English (AAE) in NLI task. First, we propose CodeSwitch, a greedy unidirectional morphosyntactically-informed rule-based translation method for data augmentation. Next, we use CodeSwitch to present a preliminary study to determine if demographic language features do in fact influence models to produce false predictions. Then, we conduct experiments on two popular datasets and propose two simple, yet effective and generalizable debiasing methods. Our findings show that NLI models (e.g. BERT) trained under our proposed frameworks outperform traditional large language models while maintaining or even improving the prediction performance. In addition, we intend to release CodeSwitch, in hopes of promoting dialectal language diversity in training data to both reduce the discriminatory societal impacts and improve model robustness of downstream NLP tasks.

Word embeddings are critical for numerous NLP tasks but their evaluation in actual under-resourced settings needs further examination. This paper presents a case study in Bribri, a Chibchan language from Costa Rica. Four experiments were adapted from English: Word similarities, WordSim353 correlations, odd-one-out tasks and analogies. Here we discuss their adaptation to an under-resourced Indigenous language and we use them to measure semantic and morphological learning. We trained 96 word2vec models with different hyperparameter combinations. The best models for this under-resourced scenario were Skip-grams with an intermediate size (100 dimensions) and large window sizes (10). These had an average correlation of r=0.28 with WordSim353, a 76% accuracy in semantic odd-one-out and 70% accuracy in structural/morphological odd-one-out. The performance was lower for the analogies: The best models could find the appropriate semantic target amongst the first 25 results approximately 60% of the times, but could only find the morphological/structural target 11% of the times. Future research needs to further explore the patterns of morphological/structural learning, to examine the behavior of deep learning embeddings, and to establish a human baseline. This project seeks to improve Bribri NLP and ultimately help in its maintenance and revitalization.

Diversity can be decomposed into three distinct concepts, namely: variety, balance and disparity. This paper borrows from the extensive formalization and measures of diversity developed in ecology in order to evaluate the variety and balance of multiword expression annotation produced by automatic annotation systems. The measures of richness, normalized richness, and two variations of Hill’s evenness are considered in this paper. We observe how these measures behave against increasingly smaller samples of gold annotations of multiword expressions and use their comportment to validate or invalidate their pertinence for multiword expressions in annotated texts. We apply the validated measures to annotations in 14 languages produced by systems during the PARSEME shared task on automatic identification of multiword expressions and on the gold versions of the corpora. We also explore the limits of such evaluation by studying the impact of lemmatization errors in the Turkish corpus used in the shared task.

The human recognition system has presented the remarkable ability to effortlessly learn novel knowledge from only a few trigger events based on prior knowledge, which is called insight learning. Mimicking such behavior on Knowledge Graph Reasoning (KGR) is an interesting and challenging research problem with many practical applications. Simultaneously, existing works, such as knowledge embedding and few-shot learning models, have been limited to conducting KGR in either “seen-to-seen” or “unseen-to-unseen” scenarios. To this end, we propose a neural insight learning framework named Eureka to bridge the “seen” to “unseen” gap. Eureka is empowered to learn the seen relations with sufficient training triples while providing the flexibility of learning unseen relations given only one trigger without sacrificing its performance on seen relations. Eureka meets our expectation of the model to acquire seen and unseen relations at no extra cost, and eliminate the need to retrain when encountering emerging unseen relations. Experimental results on two real-world datasets demonstrate that the proposed framework also outperforms various state-of-the-art baselines on datasets of both seen and unseen relations.

In many linguistic fields requiring annotated data, multiple interpretations of a single item are possible. Multi-label annotations more accurately reflect this possibility. However, allowing for multi-label annotations also affects the chance that two coders agree with each other. Calculating inter-coder agreement for multi-label datasets is therefore not trivial. In the current contribution, we evaluate different metrics for calculating agreement on multi-label annotations: agreement on the intersection of annotated labels, an augmented version of Cohen’s Kappa, and precision, recall and F1. We propose a bootstrapping method to obtain chance agreement for each measure, which allows us to obtain an adjusted agreement coefficient that is more interpretable. We demonstrate how various measures affect estimates of agreement on simulated datasets and present a case study of discourse relation annotations. We also show how the proportion of double labels, and the entropy of the label distribution, influences the measures outlined above and how a bootstrapped adjusted agreement can make agreement measures more comparable across datasets in multi-label scenarios.

The ability to generalize compositionally is key to understanding the potentially infinite number of sentences that can be constructed in a human language from only a finite number of words. Investigating whether NLP models possess this ability has been a topic of interest: SCAN (Lake and Baroni, 2018) is one task specifically proposed to test for this property. Previous work has achieved impressive empirical results using a group-equivariant neural network that naturally encodes a useful inductive bias for SCAN (Gordon et al., 2020). Inspired by this, we introduce a novel group-equivariant architecture that incorporates a group-invariant hard alignment mechanism. We find that our network’s structure allows it to develop stronger equivariance properties than existing group-equivariant approaches. We additionally find that it outperforms previous group-equivariant networks empirically on the SCAN task. Our results suggest that integrating group-equivariance into a variety of neural architectures is a potentially fruitful avenue of research, and demonstrate the value of careful analysis of the theoretical properties of such architectures.

This paper explores the task of identifying the overall sentiment expressed towards volitional entities (persons and organizations) in a document - what we refer to as Entity-Level Sentiment Analysis (ELSA). While identifying sentiment conveyed towards an entity is well researched for shorter texts like tweets, we find little to no research on this specific task for longer texts with multiple mentions and opinions towards the same entity. This lack of research would be understandable if ELSA can be derived from existing tasks and models. To assess this, we annotate a set of professional reviews for their overall sentiment towards each volitional entity in the text. We sample from data already annotated for document-level, sentence-level, and target-level sentiment in a multi-domain review corpus, and our results indicate that there is no single proxy task that provides this overall sentiment we seek for the entities at a satisfactory level of performance. We present a suite of experiments aiming to assess the contribution towards ELSA provided by document-, sentence-, and target-level sentiment analysis, and provide a discussion of their shortcomings. We show that sentiment in our dataset is expressed not only with an entity mention as target, but also towards targets with a sentiment-relevant relation to a volitional entity. In our data, these relations extend beyond anaphoric coreference resolution, and our findings call for further research of the topic. Finally, we also present a survey of previous relevant work.

Large-scale pretrained language models have led to significant improvements in Natural Language Processing. Unfortunately, they come at the cost of high computational and storage requirements that complicate their deployment on low-resource devices. This issue can be addressed by distilling knowledge from larger models to smaller ones through pseudo-labels on task-specific datasets. However, this can be difficult for tasks with very limited data. To overcome this challenge, we present a novel approach where knowledge can be distilled from a teacher model to a student model through the generation of synthetic data. For this to be done, we first fine-tune the teacher and student models, as well as a Natural Language Generation (NLG) model, on the target task dataset. We then let both student and teacher work together to condition the NLG model to generate examples that can enhance the performance of the student. We tested our approach on two data generation methods: a) Targeted generation using the Monte Carlo Tree Search (MCTS) algorithm, and b) A Non-Targeted Text Generation (NTTG) method. We evaluate the effectiveness of our approaches against a baseline that uses the BERT model for data augmentation through random word replacement. By testing this approach on the SST-2, MRPC, YELP-2, DBpedia, and TREC-6 datasets, we consistently witnessed considerable improvements over the word-replacement baseline.

Dependency parsing is an important fundamental natural language processing task which analyzes the syntactic structure of an input sentence by illustrating the syntactic relations between words. To improve dependency parsing, leveraging existing dependency parsers and extra data (e.g., through semi-supervised learning) has been demonstrated to be effective, even though the final parsers are trained on inaccurate (but massive) data. In this paper, we propose a frustratingly easy approach to improve graph-based dependency parsing, where a structure-aware encoder is pre-trained on auto-parsed data by predicting the word dependencies and then fine-tuned on gold dependency trees, which differs from the usual pre-training process that aims to predict the context words along dependency paths. Experimental results and analyses demonstrate the effectiveness and robustness of our approach to benefit from the data (even with noise) processed by different parsers, where our approach outperforms strong baselines under different settings with different dependency standards and model architectures used in pre-training and fine-tuning. More importantly, further analyses find that only 2K auto-parsed sentences are required to obtain improvement when pre-training vanilla BERT-large based parser without requiring extra parameters.

Today the pre-trained language models achieve great success for question generation (QG) task and significantly outperform traditional sequence-to-sequence approaches. However, the pre-trained models treat the input passage as a flat sequence and are thus not aware of the text structure of input passage. For QG task, we model text structure as answer position and syntactic dependency, and propose answer localness modeling and syntactic mask attention to address these limitations. Specially, we present localness modeling with a Gaussian bias to enable the model to focus on answer-surrounded context, and propose a mask attention mechanism to make the syntactic structure of input passage accessible in question generation process. Experiments on SQuAD dataset show that our proposed two modules improve performance over the strong pre-trained model ProphetNet, and combing them together achieves very competitive results with the state-of-the-art pre-trained model.

Pre-trained language models (PLMs) such as BERT and RoBERTa have dramatically improved the performance of various natural language processing tasks. Although these models are trained on large amounts of raw text, they have no explicit grounding in real-world entities. Knowledge graphs (KGs) are manually annotated with factual knowledge and store the relations between nodes corresponding to entities as labeled edges. This paper proposes a mechanism called KG-attention, which integrates the structure of a KG into recent PLM architectures. Unlike the existing PLM+KG integration methods, KG-attention generalizes the embeddings of neighboring entities using the relation embeddings; accordingly, it can handle relations between unconnected entities in the KG. Experimental results demonstrated that our method achieved significant improvements in a relation classification task, an entity typing task, and several language comprehension tasks.

Dense video captioning aims to identify the events of interest in an input video, and generate descriptive captions for each event. Previous approaches usually follow a two-stage generative process, which first proposes a segment for each event, then renders a caption for each identified segment. Recent advances in large-scale sequence generation pretraining have seen great success in unifying task formulation for a great variety of tasks, but so far, more complex tasks such as dense video captioning are not able to fully utilize this powerful paradigm. In this work, we show how to model the two subtasks of dense video captioning jointly as one sequence generation task, and simultaneously predict the events and the corresponding descriptions. Experiments on YouCook2 and ViTT show encouraging results and indicate the feasibility of training complex tasks such as end-to-end dense video captioning integrated into large-scale pretrained models.

The state of bridging resolution research is rather unsatisfactory: not only are state-of-the-art resolvers evaluated in unrealistic settings, but the neural models underlying these resolvers are weaker than those used for entity coreference resolution. In light of these problems, we evaluate bridging resolvers in an end-to-end setting, strengthen them with better encoders, and attempt to gain a better understanding of them via perturbation experiments and a manual analysis of their outputs.

Word embeddings learned using the distributional hypothesis (e.g., GloVe, Word2vec) are good at encoding various lexical-semantic relations. However, they do not capture the emotion aspects of words. We present a novel retrofitting method for updating the vectors of emotion bearing words like fun, offence, angry, etc. The retrofitted embeddings achieve better inter-cluster and intra-cluster distance for words having the same emotions, e.g., the joy cluster containing words like fun, happiness, etc., and the anger cluster with words like offence, rage, etc., as evaluated through different cluster quality metrics. For the downstream tasks on sentiment analysis and sarcasm detection, simple classification models, such as SVM and Attention Net, learned using our retrofitted embeddings perform better than their pre-trained counterparts (about 1.5 % improvement in F1-score) as well as other benchmarks. Furthermore, the difference in performance is more pronounced in the limited data setting.