EMNLP 2020 Papers

Ezafe is a grammatical particle in some Iranian languages that links two words together. Regardless of the important information it conveys, it is almost always not indicated in Persian script, resulting in mistakes in reading complex sentences and errors in natural language processing tasks. In this paper, we experiment with different machine learning methods to achieve state-of-the-art results in the task of ezafe recognition. Transformer-based methods, BERT and XLMRoBERTa, achieve the best results, the latter achieving 2.68% F1-score more than the previous state-of-the-art. We, moreover, use ezafe information to improve Persian part-of-speech tagging results and show that such information will not be useful to transformer-based methods and explain why that might be the case.

Sensor metadata tagging, akin to the named entity recognition task, provides key contextual information (e.g., measurement type and location) about sensors for running smart building applications. Unfortunately, sensor metadata in different buildings often follows distinct naming conventions. Therefore, learning a tagger currently requires extensive annotations on a per building basis. In this work, we propose a novel framework, SeNsER, which learns a sensor metadata tagger for a new building based on its raw metadata and some existing fully annotated building. It leverages the commonality between different buildings: At the character level, it employs bidirectional neural language models to capture the shared underlying patterns between two buildings and thus regularizes the feature learning process; At the word level, it leverages as features the k-mers existing in the fully annotated building. During inference, we further incorporate the information obtained from sources such as Wikipedia as prior knowledge. As a result, SeNsER shows promising results in extensive experiments on multiple real-world buildings.

Recent studies have demonstrated the effectiveness of cross-lingual language model pre-training on different NLP tasks, such as natural language inference and machine translation. In our work, we test this approach on social media data, which are particularly challenging to process within this framework, since the limited length of the textual messages and the irregularity of the language make it harder to learn meaningful encodings. More specifically, we propose a hybrid emoji-based Masked Language Model (MLM) to leverage the common information conveyed by emojis across different languages and improve the learned cross-lingual representation of short text messages, with the goal to perform zero- shot abusive language detection. We compare the results obtained with the original MLM to the ones obtained by our method, showing improved performance on German, Italian and Spanish.

Multilingual BERT (mBERT) has shown reasonable capability for zero-shot cross-lingual transfer when fine-tuned on downstream tasks. Since mBERT is not pre-trained with explicit cross-lingual supervision, transfer performance can further be improved by aligning mBERT with cross-lingual signal. Prior work propose several approaches to align contextualised embeddings. In this paper we analyse how different forms of cross-lingual supervision and various alignment methods influence the transfer capability of mBERT in zero-shot setting. Specifically, we compare parallel corpora vs dictionary-based supervision and rotational vs fine-tuning based alignment methods. We evaluate the performance of different alignment methodologies across eight languages on two tasks: Name Entity Recognition and Semantic Slot Filling. In addition, we propose a novel normalisation method which consistently improves the performance of rotation-based alignment including a notable 3% F1 improvement for distant and typologically dissimilar languages. Importantly we identify the biases of the alignment methods to the type of task and proximity to the transfer language. We also find that supervision from parallel corpus is generally superior to dictionary alignments.

We investigate a long-perceived shortcoming in the typical use of BLEU: its reliance on a single reference. Using modern neural paraphrasing techniques, we study whether automatically generating additional diverse references can provide better coverage of the space of valid translations and thereby improve its correlation with human judgments. Our experiments on the into-English language directions of the WMT19 metrics task (at both the system and sentence level) show that using paraphrased references does generally improve BLEU, and when it does, the more diverse the better. However, we also show that better results could be achieved if those paraphrases were to specifically target the parts of the space most relevant to the MT outputs being evaluated. Moreover, the gains remain slight even when human paraphrases are used, suggesting inherent limitations to BLEU’s capacity to correctly exploit multiple references. Surprisingly, we also find that adequacy appears to be less important, as shown by the high results of a strong sampling approach, which even beats human paraphrases when used with sentence-level BLEU.

Large-scale pretrained language models have achieved outstanding performance on natural language understanding tasks. However, it is still under investigating how to apply them to dialogue generation tasks, especially those with responses conditioned on multiple sources. Previous work simply concatenates all input sources or averages information from different input sources. In this work, we study dialogue models with multiple input sources adapted from the pretrained language model GPT2. We explore various methods to fuse multiple separate attention information corresponding to different sources. Our experimental results show that proper fusion methods deliver higher relevance with dialogue history than simple fusion baselines.

Form understanding depends on both textual contents and organizational structure. Although modern OCR performs well, it is still challenging to realize general form understanding because forms are commonly used and of various formats. The table detection and handcrafted features in previous works cannot apply to all forms because of their requirements on formats. Therefore, we concentrate on the most elementary components, the key-value pairs, and adopt multimodal methods to extract features. We consider the form structure as a tree-like or graph-like hierarchy of text fragments. The parent-child relation corresponds to the key-value pairs in forms. We utilize the state-of-the-art models and design targeted extraction modules to extract multimodal features from semantic contents, layout information, and visual images. A hybrid fusion method of concatenation and feature shifting is designed to fuse the heterogeneous features and provide an informative joint representation. We adopt an asymmetric algorithm and negative sampling in our model as well. We validate our method on two benchmarks, MedForm and FUNSD, and extensive experiments demonstrate the effectiveness of our method.

Despite recent success in neural task-oriented dialogue systems, developing such a real-world system involves accessing large-scale knowledge bases (KBs), which cannot be simply encoded by neural approaches, such as memory network mechanisms. To alleviate the above problem, we propose , an end-to-end trainable text-to-SQL guided framework to learn a neural agent that interacts with KBs using the generated SQL queries. Specifically, the neural agent first learns to ask and confirm the customer’s intent during the multi-turn interactions, then dynamically determining when to ground the user constraints into executable SQL queries so as to fetch relevant information from KBs. With the help of our method, the agent can use less but more accurate fetched results to generate useful responses efficiently, instead of incorporating the entire KBs. We evaluate the proposed method on the AirDialogue dataset, a large corpus released by Google, containing the conversations of customers booking flight tickets from the agent. The experimental results show that significantly improves over previous work in terms of accuracy and the BLEU score, which demonstrates not only the ability to achieve the given task but also the good quality of the generated dialogues.

Whenever researchers write a paper, the same question occurs: “Where to submit?” In this work, we introduce WTS, an open and interpretable NLP system that recommends conferences and journals to researchers based on the title, abstract, and/or keywords of a given paper. We adapt the TextCNN architecture and automatically analyze its predictions using the Integrated Gradients method to highlight words and phrases that led to the recommendation of a scientific venue. We train and test our method on publications from the fields of artificial intelligence (AI) and medicine, both derived from the Semantic Scholar dataset. WTS achieves an Accuracy@5 of approximately 83% for AI papers and 95% in the field of medicine. It is open source and available for testing on https://wheretosubmit.ml.

Dividing biomedical abstracts into several segments with rhetorical roles is essential for supporting researchers’ information access in the biomedical domain. Conventional methods have regarded the task as a sequence labeling task based on sequential sentence classification, i.e., they assign a rhetorical label to each sentence by considering the context in the abstract. However, these methods have a critical problem: they are prone to mislabel longer continuous sentences with the same rhetorical label. To tackle the problem, we propose sequential span classification that assigns a rhetorical label, not to a single sentence but to a span that consists of continuous sentences. Accordingly, we introduce Neural Semi-Markov Conditional Random Fields to assign the labels to such spans by considering all possible spans of various lengths. Experimental results obtained from PubMed 20k RCT and NICTA-PIBOSO datasets demonstrate that our proposed method achieved the best micro sentence-F1 score as well as the best micro span-F1 score.

Neural text generation (data- or text-to-text) demonstrates remarkable performance when training data is abundant which for many applications is not the case. To collect a large corpus of parallel data, heuristic rules are often used but they inevitably let noise into the data, such as phrases in the output which cannot be explained by the input. Consequently, models pick up on the noise and may hallucinate–generate fluent but unsupported text. Our contribution is a simple but powerful technique to treat such hallucinations as a controllable aspect of the generated text, without dismissing any input and without modifying the model architecture. On the WikiBio corpus (Lebret et al., 2016), a particularly noisy dataset, we demonstrate the efficacy of the technique both in an automatic and in a human evaluation.

In order to improve the sample-efficiency of deep reinforcement learning (DRL), we implemented imagination augmented agent (I2A) in spoken dialogue systems (SDS). Although I2A achieves a higher success rate than baselines by augmenting predicted future into a policy network, its complicated architecture introduces unwanted instability. In this work, we propose actor-double-critic (ADC) to improve the stability and overall performance of I2A. ADC simplifies the architecture of I2A to reduce excessive parameters and hyper-parameters. More importantly, a separate model-based critic shares parameters between actions and makes back-propagation explicit. In our experiments on Cambridge Restaurant Booking task, ADC enhances success rates considerably and shows robustness to imperfect environment models. In addition, ADC exhibits the stability and sample-efficiency as significantly reducing the baseline standard deviation of success rates and reaching the 80% success rate with half training data.

Existing NLP datasets contain various biases that models can easily exploit to achieve high performances on the corresponding evaluation sets. However, focusing on dataset-specific biases limits their ability to learn more generalizable knowledge about the task from more general data patterns. In this paper, we investigate the impact of debiasing methods for improving generalization and propose a general framework for improving the performance on both in-domain and out-of-domain datasets by concurrent modeling of multiple biases in the training data. Our framework weights each example based on the biases it contains and the strength of those biases in the training data. It then uses these weights in the training objective so that the model relies less on examples with high bias weights. We extensively evaluate our framework on extractive question answering with training data from various domains with multiple biases of different strengths. We perform the evaluations in two different settings, in which the model is trained on a single domain or multiple domains simultaneously, and show its effectiveness in both settings compared to state-of-the-art debiasing methods.

Event extraction, which aims to identify event triggers of pre-defined event types and their arguments of specific roles, is a challenging task in NLP. Most traditional approaches formulate this task as classification problems, with event types or argument roles taken as golden labels. Such approaches fail to model rich interactions among event types and arguments of different roles, and cannot generalize to new types or roles. This work proposes a new paradigm that formulates event extraction as multi-turn question answering. Our approach, MQAEE, casts the extraction task into a series of reading comprehension problems, by which it extracts triggers and arguments successively from a given sentence. A history answer embedding strategy is further adopted to model question answering history in the multi-turn process. By this new formulation, MQAEE makes full use of dependency among arguments and event types, and generalizes well to new types with new argument roles. Empirical results on ACE 2005 shows that MQAEE outperforms current state-of-the-art, pushing the final F1 of argument extraction to 53.4% (+2.0%). And it also has a good generalization ability, achieving competitive performance on 13 new event types even if trained only with a few samples of them.

The state-of-the-art Aspect-based Sentiment Analysis (ABSA) approaches are mainly based on either detecting aspect terms and their corresponding sentiment polarities, or co-extracting aspect and opinion terms. However, the extraction of aspect-sentiment pairs lacks opinion terms as a reference, while co-extraction of aspect and opinion terms would not lead to meaningful pairs without determining their sentiment dependencies. To address the issue, we present a novel view of ABSA as an opinion triplet extraction task, and propose a multi-task learning framework to jointly extract aspect terms and opinion terms, and simultaneously parses sentiment dependencies between them with a biaffine scorer. At inference phase, the extraction of triplets is facilitated by a triplet decoding method based on the above outputs. We evaluate the proposed framework on four SemEval benchmarks for ASBA. The results demonstrate that our approach significantly outperforms a range of strong baselines and state-of-the-art approaches.

We present a novel way of injecting factual knowledge about entities into the pretrained BERT model (Devlin et al., 2019): We align Wikipedia2Vec entity vectors (Yamada et al., 2016) with BERT’s native wordpiece vector space and use the aligned entity vectors as if they were wordpiece vectors. The resulting entity-enhanced version of BERT (called E-BERT) is similar in spirit to ERNIE (Zhang et al., 2019) and KnowBert (Peters et al., 2019), but it requires no expensive further pre-training of the BERT encoder. We evaluate E-BERT on unsupervised question answering (QA), supervised relation classification (RC) and entity linking (EL). On all three tasks, E-BERT outperforms BERT and other baselines. We also show quantitatively that the original BERT model is overly reliant on the surface form of entity names (e.g., guessing that someone with an Italian-sounding name speaks Italian), and that E-BERT mitigates this problem.

Neural dialogue response generation has gained much popularity in recent years. Maximum Likelihood Estimation (MLE) objective is widely adopted in existing dialogue model learning. However, models trained with MLE objective function are plagued by the low-diversity issue when it comes to the open-domain conversational setting. Inspired by the observation that humans not only learn from the positive signals but also benefit from correcting behaviors of undesirable actions, in this work, we introduce contrastive learning into dialogue generation, where the model explicitly perceives the difference between the well-chosen positive and negative utterances. Specifically, we employ a pretrained baseline model as a reference. During contrastive learning, the target dialogue model is trained to give higher conditional probabilities for the positive samples, and lower conditional probabilities for those negative samples, compared to the reference model. To manage the multi-mapping relations prevalent in human conversation, we augment contrastive dialogue learning with group-wise dual sampling. Extensive experimental results show that the proposed group-wise contrastive learning framework is suited for training a wide range of neural dialogue generation models with very favorable performance over the baseline training approaches.

One of the biggest bottlenecks in building accurate, high coverage neural open IE systems is the need for large labelled corpora. The diversity of open domain corpora and the variety of natural language expressions further exacerbate this problem. In this paper, we propose a syntactic and semantic-driven learning approach, which can learn neural open IE models without any human-labelled data by leveraging syntactic and semantic knowledge as noisier, higher-level supervision. Specifically, we first employ syntactic patterns as data labelling functions and pretrain a base model using the generated labels. Then we propose a syntactic and semantic-driven reinforcement learning algorithm, which can effectively generalize the base model to open situations with high accuracy. Experimental results show that our approach significantly outperforms the supervised counterparts, and can even achieve competitive performance to supervised state-of-the-art (SoA) model.

Dialogue state tracking (DST) aims at estimating the current dialogue state given all the preceding conversation. For multi-domain DST, the data sparsity problem is a major obstacle due to increased numbers of state candidates and dialogue lengths. To encode the dialogue context efficiently, we utilize the previous dialogue state (predicted) and the current dialogue utterance as the input for DST. To consider relations among different domain-slots, the schema graph involving prior knowledge is exploited. In this paper, a novel context and schema fusion network is proposed to encode the dialogue context and schema graph by using internal and external attention mechanisms. Experiment results show that our approach can outperform strong baselines, and the previous state-of-the-art method (SOM-DST) can also be improved by our proposed schema graph.