Submission Type

Lipschitz Constrained Parameter Initialization for Deep Transformers

Hongfei Xu, Qiuhui Liu, Josef van Genabith, Deyi Xiong and Jingyi Zhang

The Transformer translation model employs residual connection and layer normalization to ease the optimization difficulties caused by its multi-layer encoder/decoder structure. Previous research shows that even with residual connection and layer normalization, deep Transformers still have difficulty in training, and particularly Transformer models with more than 12 encoder/decoder layers fail to converge. In this paper, we first empirically demonstrate that a simple modification made in the official implementation, which changes the computation order of residual connection and layer normalization, can significantly ease the optimization of deep Transformers. We then compare the subtle differences in computation order in considerable detail, and present a parameter initialization method that leverages the Lipschitz constraint on the initialization of Transformer parameters that effectively ensures training convergence. In contrast to findings in previous research we further demonstrate that with Lipschitz parameter initialization, deep Transformers with the original computation order can converge, and obtain significant BLEU improvements with up to 24 layers. In contrast to previous research which focuses on deep encoders, our approach additionally enables Transformers to also benefit from deep decoders.

Fine-grained Fact Verification with Kernel Graph Attention Network

Zhenghao Liu, Chenyan Xiong, Maosong Sun and Zhiyuan Liu

Fact Verification requires fine-grained natural language inference capability that finds subtle clues to identify the syntactical and semantically correct but not well-supported claims. This paper presents Kernel Graph Attention Network (KGAT), which conducts more fine-grained fact verification with kernel-based attentions. Given a claim and a set of potential evidence sentences that form an evidence graph, KGAT introduces node kernels, which better measure the importance of the evidence node, and edge kernels, which conduct fine-grained evidence propagation in the graph, into Graph Attention Networks for more accurate fact verification. KGAT achieves a 70.38% FEVER score and significantly outperforms existing fact verification models on FEVER, a large-scale benchmark for fact verification. Our analyses illustrate that, compared to dot-product attentions, the kernel-based attention concentrates more on relevant evidence sentences and meaningful clues in the evidence graph, which is the main source of KGAT's effectiveness. All source codes of this work are available at https://github.com/thunlp/KernelGAT.

Towards Better Non-Tree Argument Mining: Proposition-Level Biaffine Parsing with Task-Specific Parameterization

Gaku Morio, Hiroaki Ozaki, Terufumi Morishita, Yuta Koreeda and Kohsuke Yanai

State-of-the-art argument mining studies have advanced the techniques for predicting argument structures. However, the technology for capturing non-tree-structured arguments is still in its infancy.
In this paper, we focus on non-tree argument mining with a neural model. We jointly predict proposition types and edges between propositions. Our proposed model incorporates (i) task-specific parameterization (TSP) that effectively encodes a sequence of propositions and (ii) a proposition-level biaffine attention (PLBA) that can predict a non-tree argument consisting of edges. Experimental results show that both TSP and PLBA boost edge prediction performance compared to baselines.

Learning Source Phrase Representations for Neural Machine Translation

Hongfei Xu, Josef van Genabith, Deyi Xiong, Qiuhui Liu and Jingyi Zhang

The Transformer translation model (Vaswani et al., 2017) based on a multi-head attention mechanism can be computed effectively in parallel and has significantly pushed forward the performance of Neural Machine Translation (NMT). Though intuitively the attentional network can connect distant words via shorter network paths than RNNs, empirical analysis demonstrates that it still has difficulty in fully capturing long-distance dependencies (Tang et al., 2018). Considering that modeling phrases instead of words has significantly improved the Statistical Machine Translation (SMT) approach through the use of larger translation blocks ("phrases") and its reordering ability, modeling NMT at phrase level is an intuitive proposal to help the model capture long-distance relationships. In this paper, we first propose an attentive phrase representation generation mechanism which is able to generate phrase representations from corresponding token representations. In addition, we incorporate the generated phrase representations into the Transformer translation model to enhance its ability to capture long-distance relationships. In our experiments, we obtain significant improvements on the WMT 14 English-German and English-French tasks on top of the strong Transformer baseline, which shows the effectiveness of our approach. Our approach helps Transformer Base models perform at the level of Transformer Big models, and even significantly better for long sentences, but with substantially fewer parameters and training steps. The fact that phrase representations help even in the big setting further supports our conjecture that they make a valuable contribution to long-distance relations.

What Question Answering can Learn from Trivia Nerds

Jordan Boyd-Graber and Benjamin Börschinger

In addition to the traditional task of machines answering questions, question answering (QA) research creates interesting, challenging questions that help systems how to answer questions and reveal the best systems. We argue that creating a QA dataset—and the ubiquitous leaderboard that goes with it—closely resembles running a trivia tournament: you write questions, have agents (either humans or machines) answer the questions, and declare a winner. However, the research community has ignored the hard-learned lessons from decades of the trivia community creating vibrant, fair, and effective question answering competitions. After detailing problems with existing QA datasets, we outline the key lessons—removing ambiguity, discriminating skill, and adjudicating disputes---that can transfer to QA research and how they might be implemented.

Cross-media Structured Common Space for Multimedia Event Extraction

Manling Li, Alireza Zareian, Qi Zeng, Spencer Whitehead, Di Lu, Heng Ji and Shih-Fu Chang

We introduce a new task, MultiMedia Event Extraction, which aims to extract events and their arguments from multimedia documents. We develop the first benchmark and collect a dataset of 245 multimedia news articles with extensively annotated events and arguments. We propose a novel method, Weakly Aligned Structured Embedding (WASE), that encodes structured representations of semantic information from textual and visual data into a common embedding space. The structures are aligned across modalities by employing a weakly supervised training strategy, which enables exploiting available resources without explicit cross-media annotation. Compared to uni-modal state-of-the-art methods, our approach achieves 4.0% and 9.8% absolute F-score gains on text event argument role labeling and visual event extraction. Compared to state-of-the-art multimedia unstructured representations, we achieve 8.3% and 5.0% absolute F-score gains on multimedia event extraction and argument role labeling, respectively. By utilizing images, we extract 21.4% more event mentions than traditional text-only methods.

Good-Enough Compositional Data Augmentation

We propose a simple data augmentation protocol aimed at providing a compositional inductive bias in conditional and unconditional sequence models. Under this protocol, synthetic training examples are constructed by taking real training examples and replacing (possibly discontinuous) fragments with other fragments that appear in at least one similar environment. The protocol is model-agnostic and useful for a variety of tasks. Applied to neural sequence-to-sequence models, it reduces error rate by as much as 87% on diagnostic tasks from the SCAN dataset and 16% on a semantic parsing task. Applied to n-gram language models, it reduces perplexity by roughly 1% on small corpora in several languages.

Facet-Aware Evaluation for Extractive Summarization

Yuning Mao, Liyuan Liu, Qi Zhu, Xiang Ren and Jiawei Han

Commonly adopted metrics for extractive summarization focus on lexical overlap at the token level. In this paper, we present a facet-aware evaluation setup for better assessment of the information coverage in extracted summaries. Specifically, we treat each sentence in the reference summary as a \textit{facet}, identify the sentences in the document that express the semantics of each facet as \textit{support sentences} of the facet, and automatically evaluate extractive summarization methods by comparing the indices of extracted sentences and support sentences of all the facets in the reference summary. To facilitate this new evaluation setup, we construct an extractive version of the CNN/Daily Mail dataset and perform a thorough quantitative investigation, through which we demonstrate that facet-aware evaluation manifests better correlation with human judgment than ROUGE, enables fine-grained evaluation as well as comparative analysis, and reveals valuable insights of state-of-the-art summarization methods.
Data can be found at \url{https://github.com/morningmoni/FAR}.

A Unified MRC Framework for Named Entity Recognition

Xiaoya Li, Jingrong Feng, Yuxian Meng, Qinghong Han, Fei Wu and Jiwei Li

The task of named entity recognition (NER) is normally divided into nested NER and flat NER depending on whether named entities are nested or not.Models are usually separately developed for the two tasks, since sequence labeling models, the most widely used backbone for flat NER, are only able to assign a single label to a particular token, which is unsuitable for nested NER where a token may be assigned several labels.

In this paper, we propose a unified framework that is capable of handling both flat and nested NER tasks. Instead of treating the task of NER as a sequence labeling problem, we propose to formulate it as a machine reading comprehension (MRC) task. For example, extracting entities with the \textsc{per} label is formalized as extracting answer spans to the question ``{\it which person is mentioned in the text}".This formulation naturally tackles the entity overlapping issue in nested NER: the extraction of two overlapping entities with different categories requires answering two independent questions. Additionally, since the query encodes informative prior knowledge, this strategy facilitates the process of entity extraction, leading to better performances for not only nested NER, but flat NER.

We conduct experiments on both nested and flat NER datasets.Experiment results demonstrate the effectiveness of the proposed formulation. We are able to achieve a vast amount of performance boost over current SOTA models on nested NER datasets, i.e., +1.28, +2.55, +5.44, +6.37,respectively on ACE04, ACE05, GENIA and KBP17, along with SOTA results on flat NER datasets, i.e., +0.24, +1.95, +0.21, +1.49 respectively on English CoNLL 2003, English OntoNotes 5.0, Chinese MSRA and Chinese OntoNotes 4.0.

Beyond User Self-Reported Likert Scale Ratings: A Comparison Model for Automatic Dialog Evaluation

Weixin Liang, James Zou and Zhou Yu

Open Domain dialog system evaluation is one of the most important challenges in dialog research. Existing automatic evaluation metrics, such as BLEU are mostly reference-based. They calculate the difference between the generated response and a limited number of available references. Likert-score based self-reported user rating is widely adopted by social conversational systems, such as Amazon Alexa Prize chatbots. However, self-reported user rating suffers from bias and variance among different users. To alleviate this problem, we formulate dialog evaluation as a comparison task. We also propose an automatic evaluation model CMADE (Comparison Model for Automatic Dialog Evaluation) that automatically cleans self-reported user ratings as it trains on them. Specifically, we first use a self-supervised method to learn better dialog feature representation, and then use KNN and Shapley to remove confusing samples. Our experiments show that CMADE achieves 89.2% accuracy in the dialog comparison task.

Negative Training for Neural Dialogue Response Generation

Tianxing He and James Glass

Although deep learning models have brought tremendous advancements to the field of open-domain dialogue response generation, recent research results have revealed that the trained models have undesirable generation behaviors, such as malicious responses and generic (boring) responses. In this work, we propose a framework named ``Negative Training" to minimize such behaviors. Given a trained model, the framework will first find generated samples that exhibit the undesirable behavior, and then use them to feed negative training signals for fine-tuning the model. Our experiments show that negative training can significantly reduce the hit rate of malicious responses, or discourage frequent responses and improve response diversity.

Dialogue-Based Relation Extraction

Dian Yu, Kai Sun, Claire Cardie and Dong Yu

We present the first human-annotated dialogue-based relation extraction (RE) dataset DialogRE, aiming to support the prediction of relation(s) between two arguments that appear in a dialogue. We further offer DialogRE as a platform for studying cross-sentence RE as most facts span multiple sentences. We argue that speaker-related information plays a critical role in the proposed task, based on an analysis of similarities and differences between dialogue-based and traditional RE tasks. Considering the timeliness of communication in a dialogue, we design a new metric to evaluate the performance of RE methods in a conversational setting and investigate the performance of several representative RE methods on DialogRE. Experimental results demonstrate that a speaker-aware extension on the best-performing model leads to gains in both the standard and conversational evaluation settings. DialogRE is available at https://dataset.org/dialogre/.

You Impress Me: Dialogue Generation via Mutual Persona Perception

Qian Liu, Yihong Chen, Bei Chen, Jian-Guang Lou, Zixuan Chen, Bin Zhou and Dongmei Zhang

Despite the continuing efforts to improve the engagingness and consistency of chit-chat dialogue systems, the majority of current work simply focus on mimicking human-like responses, leaving understudied the aspects of modeling understanding between interlocutors. The research in cognitive science, instead, suggests that understanding is an essential signal for a high-quality chit-chat conversation. Motivated by this, we propose P^2 Bot, a transmitter-receiver based framework with the aim of explicitly modeling understanding. Specifically, P^2 Bot incorporates mutual persona perception to enhance the quality of personalized dialogue generation. Experiments on a large public dataset, Persona-Chat, demonstrate the effectiveness of our approach, with a considerable boost over the state-of-the-art baselines across both automatic metrics and human evaluations.

Dice Loss for Data-imbalanced NLP Tasks

Xiaoya Li, Xiaofei Sun, Yuxian Meng, Junjun Liang, Fei Wu and Jiwei Li

Many NLP tasks such as tagging and machine reading comprehension are faced with the severe data imbalance issue: negative examples significantly outnumber positive examples, and the huge number of easy-negative examples overwhelms the training.
The most commonly used cross entropy (CE) criteria is actually an accuracy-oriented objective, and thus creates a discrepancy between training and test: at training time, each training instance contributes equally to the objective function, while at test time F1 score concerns more about positive examples.

In this paper, we propose to use dice loss in replacement of the standard cross-entropy objective for data-imbalanced NLP tasks. Dice loss is based on the Sørensen–Dice coefficient or Tversky index , which attaches similar importance to false positives and false negatives, and is more immune to the data-imbalance issue. To further alleviate the dominating influence from easy-negative examples in training, we propose to associate training examples with dynamically adjusted weights to deemphasize easy-negative examples. Theoretical analysis shows that this strategy narrows down the gap between the F1 score in evaluation and the dice loss in training.

With the proposed training objective, we observe significant performance boost on a wide range of data imbalanced NLP tasks. Notably, we are able to achieve SOTA results on CTB5, CTB6 and UD1.4 for the part of speech tagging task; SOTA results on CoNLL03, OntoNotes5.0, MSRA and OntoNotes4.0 for the named entity recognition task; along with competitive results on the tasks of machine reading comprehension and paraphrase identification.

Roles and Utilization of Attention Heads in Transformer-based Neural Language Models

Jae-young Jo and Sung-Hyon Myaeng

Sentence encoders based on the transformer architecture have shown promising results on various natural language tasks. The main impetus lies in the pre-trained neural language models that capture long-range dependencies among words, owing to multi-head attention that is unique in the architecture. However, little is known for how linguistic properties are processed, represented, and utilized for downstream tasks among hundreds of attention heads inside the pre-trained transformer-based model. For the initial goal of examining the roles of attention heads in handling a set of linguistic features, we conducted a set of experiments with ten probing tasks and three downstream tasks on four pre-trained transformer families (GPT, GPT2, BERT, and ELECTRA). Meaningful insights are shown through the lens of heat map visualization and utilized to propose a relatively simple sentence representation method that takes advantage of most influential attention heads, resulting in additional performance improvements on the downstream tasks.

Multi-Domain Neural Machine Translation with Word-Level Adaptive Layer-wise Domain Mixing

Haoming Jiang, Chen Liang, Chong Wang and Tuo Zhao

Many multi-domain neural machine translation (NMT) models achieve knowledge transfer by enforcing one encoder to learn shared embedding across domains. However, this design lacks adaptation to individual domains. To overcome this limitation, we propose a novel multi-domain NMT model using individual modules for each domain, on which we apply word-level, adaptive and layer-wise domain mixing. We first observe that words in a sentence are often related to multiple domains. Hence, we assume each word has a domain proportion, which indicates its domain preference. Then word representations are obtained by mixing their embedding in individual domains based on their domain proportions. We show this can be achieved by carefully designing multi-head dot-product attention modules for different domains, and eventually taking weighted averages of their parameters by word-level layer-wise domain proportions. Through this, we can achieve effective domain knowledge sharing and capture fine-grained domain-specific knowledge as well. Our experiments show that our proposed model outperforms existing ones in several NMT tasks.

Towards Debiasing Sentence Representations

Paul Pu Liang, Irene Mengze Li, Emily Zheng, Yao Chong Lim, Ruslan Salakhutdinov and Louis-Philippe Morency

As natural language processing methods are increasingly deployed in real-world scenarios such as healthcare, legal systems, and social science, it becomes necessary to recognize the role they potentially play in shaping social biases and stereotypes. Previous work has revealed the presence of social biases in widely used word embeddings involving gender, race, religion, and other social constructs. While some methods were proposed to debias these word-level embeddings, there is a need to perform debiasing at the sentence-level given the recent shift towards new contextualized sentence representations such as ELMo and BERT. In this paper, we investigate the presence of social biases in sentence-level representations and propose a new method, Sent-Debias, to reduce these biases. We show that Sent-Debias is effective in removing biases, and at the same time, preserves performance on sentence-level downstream tasks such as sentiment analysis, linguistic acceptability, and natural language understanding. We hope that our work will inspire future research on characterizing and removing social biases from widely adopted sentence representations for fairer NLP.

Predicting Degrees of Technicality in Automatic Terminology Extraction

Anna Hätty, Dominik Schlechtweg, Michael Dorna and Sabine Schulte im Walde

While automatic term extraction is a well-researched area, computational approaches to distinguish between degrees of technicality are still understudied. We semi-automatically create a German gold standard of technicality across four domains, and illustrate the impact of a web-crawled general-language corpus on technicality prediction. When defining a classification approach that combines general-language and domain-specific word embeddings, we go beyond previous work and align vector spaces to gain comparative embeddings. We suggest two novel models to exploit general- vs. domain-specific comparisons: a simple neural network model with pre-computed comparative-embedding information as input, and a multi-channel model computing the comparison internally. Both models outperform previous approaches, with the multi-channel model performing best.

A Three-Parameter Rank-Frequency Relation in Natural Languages

Chenchen Ding, Masao Utiyama and Eiichiro Sumita

We present that, the rank-frequency relation in textual data follows $f \propto r^{-\alpha}(r+\gamma)^{-\beta}$, where $f$ is the token frequency and $r$ is the rank by frequency, with ($\alpha$, $\beta$, $\gamma$) as parameters. The formulation is derived based on the empirical observation that $d^2 (x+y)/dx^2$ is a typical impulse function, where $(x,y)=(\log r, \log f)$. The formulation is the power law when \mbox{$\beta=0$} and the Zipf--Mandelbrot law when \mbox{$\alpha=0$}. We illustrate that $\alpha$ is related to the analytic features of syntax and $\beta+\gamma$ to those of morphology in natural languages from an investigation of multilingual corpora.

Automatic Generation of Citation Texts in Scholarly Papers: A Pilot Study

Xinyu Xing, Xiaosheng Fan and Xiaojun Wan

In this paper, we study the challenging problem of automatic generation of citation texts in scholarly papers. Given the context of a citing paper A and a cited paper B, the task aims to generate a short text to describe B in the given context of A. One big challenge for addressing this task is the lack of training data. Usually, explicit citation texts are easy to extract, but it is not easy to extract implicit citation texts from scholarly papers. We thus first train an implicit citation extraction model based on BERT and leverage the model to construct a large training dataset for the citation text generation task. Then we propose and train a multi-source pointer-generator network with cross attention mechanism for citation text generation. Empirical evaluation results on a manually labeled test dataset verify the efficacy of our model. This pilot study confirms the feasibility of automatically generating citation texts in scholarly papers and the technique has the great potential to help researchers prepare their scientific papers.

CorefQA: Coreference Resolution as Query-based Span Prediction

Wei Wu, Fei Wang, Arianna Yuan, Fei Wu and Jiwei Li

In this paper, we present CorefQA, an accurate and extensible approach for the coreference resolution task.
We formulate the problem as a span prediction task, like in question answering:
A query is generated for each candidate mention using its surrounding context, and a span prediction module is employed to extract the text spans of the coreferences within the document using the generated query.
This formulation comes with the following key advantages:
(1) The span prediction strategy provides the flexibility of retrieving mentions left out at the mention proposal stage;
(2) In the question answering framework, encoding the mention and its context explicitly in a query makes it possible to have a deep and thorough examination of cues embedded in the context of coreferent mentions; and
(3) A plethora of existing question answering datasets can be used for data augmentation to improve the model's generalization capability.
Experiments demonstrate significant performance boost over previous models, with 83.1 (+3.5) F1 score on the CoNLL-2012 benchmark and 87.5 (+2.5) F1 score on the GAP benchmark.

Relational Graph Attention Network for Aspect-based Sentiment Analysis

Kai Wang, Weizhou Shen, Yunyi Yang, Xiaojun Quan and Rui Wang

Aspect-based sentiment analysis aims to determine the sentiment polarity towards a specific aspect in online reviews. Most recent efforts adopt attention-based neural network models to implicitly connect aspects with opinion words. However, due to the complexity of language and the existence of multiple aspects in a single sentence, these models often confuse the connections. In this paper, we address this problem by means of effective encoding of syntax information. Firstly, we define a unified aspect-oriented dependency tree structure rooted at a target aspect by reshaping and pruning an ordinary dependency parse tree. Then, we propose a relational graph attention network (R-GAT) to encode the new tree structure for sentiment prediction. Extensive experiments are conducted on the SemEval 2014 and Twitter datasets, and the experimental results confirm that the connections between aspects and opinion words can be better established with our approach, and the performance of the graph attention network (GAT) is significantly improved as a consequence.

Reasoning with Multimodal Sarcastic Tweets via Modeling Cross-Modality Contrast and Semantic Association

Nan Xu, Zhixiong Zeng and Wenji Mao

Sarcasm is a sophisticated linguistic phenomenon to express the opposite of what one really means. With the rapid growth of social media, multimodal sarcastic tweets are widely posted on various social platforms. In multimodal context, sarcasm is no longer a pure linguistic phenomenon, and due to the nature of social media short text, the opposite is more often manifested via cross-modality expressions. Thus traditional text-based methods are insufficient to detect multimodal sarcasm. To reason with multimodal sarcastic tweets, in this paper, we propose a novel method for modeling cross-modality contrast in the associated context. Our method models both cross-modality contrast and semantic association by constructing the Decomposition and Relation Network (namely D&R Net). The decomposition network represents the commonality and discrepancy between image and text, and the relation network models the semantic association in cross-modality context. Experimental results on a public dataset demonstrate the effectiveness of our model in multimodal sarcasm detection.

SCDE: Sentence Cloze Dataset with High Quality Distractors From Examinations

Xiang Kong, Varun Gangal and Eduard Hovy

We introduce SCDE, a dataset to evaluate the performance of computational models through sentence prediction. SCDE is a human created sentence cloze dataset, collected from public school English examinations. Our task requires a model to fill up multiple blanks in a passage from a shared candidate set with distractors designed by English teachers. Experimental results demonstrate that this task requires the use of non-local, discourse-level context beyond the immediate sentence neighborhood. The blanks require joint solving and significantly impair each other’s context. Furthermore, through ablations, we show that the distractors are of high quality and make the task more challenging. Our experiments show that there is a significant performance gap between advanced models (72%) and humans (87%), encouraging future models to bridge this gap.

SpellGCN: Incorporating Phonological and Visual Similarities into Language Models for Chinese Spelling Check

Xingyi Cheng, Weidi Xu, Kunlong Chen, Shaohua Jiang, Feng Wang, Taifeng Wang, Wei Chu and Yuan Qi

Chinese Spelling Check (CSC) is a task to detect and correct spelling errors in Chinese natural language. Existing methods have made attempts to incorporate the similarity knowledge between Chinese characters. However, they take the similarity knowledge as either an external input resource or just heuristic rules. This paper proposes to incorporate phonological and visual similarity knowledge into language models for CSC via a specialized graph convolutional network (SpellGCN). The model builds a graph over the characters, and SpellGCN is learned to map this graph into a set of inter-dependent character classifiers. These classifiers are applied to the representations extracted by another network, such as BERT, enabling the whole network to be end-to-end trainable. Experiments are conducted on three human-annotated datasets. Our method achieves superior performance against previous models by a large margin.

HAT: Hardware-Aware Transformers for Efficient Natural Language Processing

Hanrui Wang, Zhanghao Wu, Zhijian Liu, Han Cai, Ligeng Zhu, Chuang Gan and Song Han

Transformers are ubiquitous in Natural Language Processing (NLP) tasks, but they are difficult to be deployed on hardware due to the intensive computation. To enable low-latency inference on resource-constrained hardware platforms, we propose to design Hardware-Aware Transformers (HAT) with neural architecture search. We first construct a large design space with arbitrary encoder-decoder attention and heterogeneous layers. Then we train a SuperTransformer that covers all candidates in the design space, and efficiently produces many SubTransformers with weight sharing. Finally, we perform an evolutionary search with a hardware latency constraint to find a specialized SubTransformer dedicated to run fast on the target hardware. Extensive experiments on four machine translation tasks demonstrate that HAT can discover efficient models for different hardware (CPU, GPU, IoT device). When running WMT’14 translation task on Raspberry Pi-4, HAT can achieve 3× speedup, 3.7× smaller size over baseline Transformer; 2.7× speedup, 3.6× smaller size over Evolved Transformer with 12,041× less search cost and no performance loss. HAT is open-sourced at https://github.com/mit-han-lab/hardware-aware-transformers.

ESPRIT: Explaining Solutions to Physical Reasoning Tasks

Nazneen Fatema Rajani, Rui Zhang, Yi Chern Tan, Stephan Zheng, Jeremy Weiss, Aadit Vyas, Abhijit Gupta, Caiming Xiong, Richard Socher and Dragomir Radev

Neural networks lack the ability to reason about qualitative physics and so cannot generalize to scenarios and tasks unseen during training. We propose ESPRIT, a framework for commonsense reasoning about qualitative physics in natural language that generates interpretable descriptions of physical events.
We use a two-step approach of first identifying the pivotal physical events in an environment and then generating natural language descriptions of those events using a data-to-text approach.
Our framework learns to generate explanations of how the physical simulation will causally evolve so that an agent or a human can easily reason about a solution using those interpretable descriptions. Human evaluations indicate that ESPRIT produces crucial fine-grained details and has high coverage of physical concepts compared to even human annotations. Dataset, code and documentation are available at \url{https://github.com/salesforce/esprit}.

Improving Adversarial Text Generation by Modeling the Distant Future

Ruiyi Zhang, Changyou Chen, Zhe Gan, Wenlin Wang, Dinghan Shen, Guoyin Wang, Zheng Wen and Lawrence Carin

Auto-regressive text generation models usually focus on local fluency, and may cause inconsistent semantic meaning in long text generation. Further, automatically generating words with similar semantics is challenging, and hand-crafted linguistic rules are difficult to apply. We consider a text planning scheme and present a model-based imitation-learning approach to alleviate the aforementioned issues. Specifically, we propose a novel guider network to focus on the generative process over a longer horizon, which can assist next-word prediction and provide intermediate rewards for generator optimization. Extensive experiments demonstrate that the proposed method leads to improved performance.

Low-Resource Generation of Multi-hop Reasoning Questions

Jianxing Yu, Wei Liu, Shuang Qiu, Qinliang Su, Kai Wang, Xiaojun Quan and Jian Yin

This paper focuses on generating multi-hop reasoning questions from the raw text in a low resource circumstance. Such questions have to be syntactically valid and need to logically correlate with the answers by deducing over multiple relations on several sentences in the text. Specifically, we first build a multi-hop generation model and guide it to satisfy the logical rationality by the reasoning chain extracted from a given text. Since the labeled data is limited and insufficient for training, we propose to learn the model with the help of a large scale of unlabeled data that is much easier to obtain. Such data contains rich expressive forms of the questions with structural patterns on syntax and semantics. These patterns can be estimated by the neural hidden semi-Markov model using latent variables. With latent patterns as a prior, we can regularize the generation model and produce the optimal results. Experimental results on the HotpotQA data set demonstrate the effectiveness of our model. Moreover, we apply the generated results to the task of machine reading comprehension and achieve significant performance improvements.

Logical Natural Language Generation from Open-Domain Tables

Wenhu Chen, Jianshu Chen, Yu Su, Zhiyu Chen and William Yang Wang

Neural natural language generation (NLG) models have recently shown remarkable progress in fluency and coherence. However, existing studies on neural NLG are primarily focused on surface-level realizations with limited emphasis on logical inference, an important aspect of human thinking and language. In this paper, we suggest a new NLG task where a model is tasked with generating natural language statements that can be \emph{logically entailed} by the facts in an open-domain semi-structured table. To facilitate the study of the proposed logical NLG problem, we use the existing TabFact dataset~\cite{chen2019tabfact} featured with a wide range of logical/symbolic inferences as our testbed, and propose new automatic metrics to evaluate the fidelity of generation models w.r.t.\ logical inference. The new task poses challenges to the existing monotonic generation frameworks due to the mismatch between sequence order and logical order. In our experiments, we comprehensively survey different generation architectures (LSTM, Transformer, Pre-Trained LM) trained with different algorithms (RL, Adversarial Training, Coarse-to-Fine) on the dataset and made following observations: 1) Pre-Trained LM can significantly boost both the fluency and logical fidelity metrics, 2) RL and Adversarial Training are trading fluency for fidelity, 3) Coarse-to-Fine generation can help partially alleviate the fidelity issue while maintaining high language fluency. The code and data are available at \url{https://github.com/wenhuchen/LogicNLG}.

BabyWalk: Going Farther in Vision-and-Language Navigation by Taking Baby Steps

Wang Zhu, Hexiang Hu, Jiacheng Chen, Zhiwei Deng, Vihan Jain, Eugene Ie and Fei Sha

Learning to follow instructions is of fundamental importance to autonomous agents for vision-and-language navigation (VLN). In this paper, we study how an agent can navigate long paths when learning from a corpus that consists of shorter ones. We show that existing state-of-the-art agents do not generalize well. To this end, we propose BabyWalk, a new VLN agent that is learned to navigate by decomposing long instructions into shorter ones (BabySteps) and completing them sequentially. A special design memory buffer is used by the agent to turn its past experiences into contexts for future steps. The learning process is composed of two phases. In the first phase, the agent uses imitation learning from demonstration to accomplish BabySteps. In the second phase, the agent uses curriculum-based reinforcement learning to maximize rewards on navigation tasks with increasingly longer instructions. We create two new benchmark datasets (of long navigation tasks) and use them in conjunction with existing ones to examine BabyWalk's generalization ability. Empirical results show that BabyWalk achieves state-of-the-art results on several metrics, in particular, is able to follow long instructions better. The codes and the datasets are released on our project page: https://github.com/Sha-Lab/babywalk.

Graph Neural News Recommendation with Unsupervised Preference Disentanglement

Linmei Hu, Siyong Xu, Chen Li, Cheng Yang, Chuan Shi, Nan Duan, Xing Xie and Ming Zhou

With the explosion of news information, personalized news recommendation has become very important for users to quickly find their interested contents. Most existing methods usually learn the representations of users and news from news contents for recommendation. However, they seldom consider high-order connectivity underlying the user-news interactions. Moreover, existing methods failed to disentangle a user's latent preference factors which cause her clicks on different news. In this paper, we model the user-news interactions as a bipartite graph and propose a novel Graph Neural News Recommendation model with Unsupervised Preference Disentanglement, named GNUD. Our model can encode high-order relationships into user and news representations by information propagation along the graph. Furthermore, the learned representations are disentangled with latent preference factors by a neighborhood routing algorithm, which can enhance expressiveness and interpretability. A preference regularizer is also designed to force each disentangled subspace to independently reflect an isolated preference, improving the quality of the disentangled representations. Experimental results on real-world news datasets demonstrate that our proposed model can effectively improve the performance of news recommendation and outperform state-of-the-art news recommendation methods.

SpanMlt: A Span-based Multi-Task Learning Framework for Pair-wise Aspect and Opinion Terms Extraction

He Zhao, Longtao Huang, Rong Zhang, Quan Lu and Hui Xue

Aspect terms extraction and opinion terms extraction are two key problems of fine-grained Aspect Based Sentiment Analysis (ABSA). The aspect-opinion pairs can provide a global profile about a product or service for consumers and opinion mining systems. However, traditional methods can not directly output aspect-opinion pairs without given aspect terms or opinion terms. Although some recent co-extraction methods have been proposed to extract both terms jointly, they fail to extract them as pairs. To this end, this paper proposes an end-to-end method to solve the task of Pair-wise Aspect and Opinion Terms Extraction (PAOTE). Furthermore, this paper treats the problem from a perspective of joint term and relation extraction rather than under the sequence tagging formulation performed in most prior works. We propose a multi-task learning framework based on shared spans, where the terms are extracted under the supervision of span boundaries. Meanwhile, the pair-wise relations are jointly identified using the span representations. Extensive experiments show that our model consistently outperforms state-of-the-art methods.

ReInceptionE: Relation-Aware Inception Network with Joint Local-Global Structural Information for Knowledge Graph Embedding

Zhiwen Xie, Guangyou Zhou, Jin Liu and Jimmy Xiangji Huang

The goal of Knowledge graph embedding (KGE) is to learn how to represent the low dimensional vectors for entities and relations based on the observed triples. The conventional shallow models are limited to their expressiveness. ConvE (Dettmers et al., 2018) takes advantage of CNN and improves the expressive power with parameter efficient operators by increasing the interactions between head and relation embeddings. However, there is no structural information in the embedding space of ConvE, and the performance is still limited by the number of interactions. The recent KBGAT (Nathani et al., 2019) provides another way to learn embeddings by adaptively utilizing structural information. In this paper, we take the benefits of ConvE and KBGAT together and propose a Relation-aware Inception network with joint local-global structural information for knowledge graph Embedding (ReInceptionE). Specifically, we first explore the Inception network to learn query embedding, which aims to further increase the interactions between head and relation embeddings. Then, we propose to use a relation-aware attention mechanism to enrich the query embedding with the local neighborhood and global entity information. Experimental results on both WN18RR and FB15k-237 datasets demonstrate that ReInceptionE achieves competitive performance compared with state-of-the-art methods.

Jointly Masked Sequence-to-Sequence Model for Non-Autoregressive Neural Machine Translation

Junliang Guo, Linli Xu and Enhong Chen

The masked language model has received remarkable attention due to its effectiveness on various natural language processing tasks. However, few works have adopted this technique in the sequence-to-sequence models. In this work, we introduce a jointly masked sequence-to-sequence model and explore its application on non-autoregressive neural machine translation~(NAT).
Specifically, we first empirically study the functionalities of the encoder and the decoder in NAT models, and find that the encoder
takes a more important role than the decoder regarding the translation quality. Therefore, we propose to train the encoder more rigorously by masking the encoder input while training.
As for the decoder, we propose to train it based on the consecutive masking of the decoder input with an $n$-gram loss function
to alleviate the problem of translating duplicate words.
The two types of masks are applied to the model jointly at the training stage.
We conduct experiments on five benchmark machine translation tasks,
and our model can achieve $27.69$/$32.24$ BLEU scores on WMT14 English-German/German-English tasks with $5+$ times speed up compared with an autoregressive model.

MLQA: Evaluating Cross-lingual Extractive Question Answering

Patrick Lewis, Barlas Oguz, Ruty Rinott, Sebastian Riedel and Holger Schwenk

Question answering (QA) models have shown rapid progress enabled by the availability of large, high-quality benchmark datasets.
Such annotated datasets are difficult and costly to collect, and rarely exist in languages other than English, making building QA systems that work well in other languages challenging.
In order to develop such systems, it is crucial to invest in high quality multilingual evaluation benchmarks to measure progress.
We present MLQA, a multi-way aligned extractive QA evaluation benchmark intended to spur research in this area. MLQA contains QA
instances in 7 languages, English, Arabic, German, Spanish, Hindi, Vietnamese and Simplified Chinese.
MLQA has over 12K instances in English and 5K in each other language, with each instance parallel between 4 languages on average.
We evaluate state-of-the-art cross-lingual models and machine-translation-based baselines
on MLQA. In all cases, transfer results are shown to be significantly behind training-language performance.

Bridging Anaphora Resolution as Question Answering

Most previous studies on bridging anaphora resolution (Poesio et al., 2004; Hou et al.,
2013b; Hou, 2018a) use the pairwise model to tackle the problem and assume that
the gold mention information is given. In this paper, we cast bridging anaphora resolution as question answering based on context. This allows us to find the antecedent for a given anaphor without knowing any gold mention information (except the anaphor itself). We present a question answering framework (BARQA) for this task, which leverages the power of transfer learning. Furthermore, we propose a novel method to generate a large amount of “quasi-bridging” training data. We show that our model pre-trained on this dataset and fine-tuned on a small amount of in-domain dataset achieves new state-of-the-art results for bridging anaphora resolution on two bridging corpora (ISNotes (Markert et al., 2012) and BASHI (Ro ̈siger, 2018)).

Learning Robust Models for e-Commerce Product Search

Thanh Nguyen, Nikhil Rao and Karthik Subbian

Showing items that do not match search query intent degrades customer experience in e-commerce. These mismatches result from counterfactual biases of the ranking algorithms toward noisy behavioral signals such as clicks and purchases in the search logs. Mitigating the problem requires a large labeled dataset, which is expensive and time-consuming to obtain. In this paper, we develop a deep, end-to-end model that learns to effectively classify mismatches and to generate hard mismatched examples to improve the classifier. We train the model end-to-end by introducing a latent variable into the cross-entropy loss that alternates between using the real and generated samples. This not only makes the classifier more robust but also boosts the overall ranking performance. Our model achieves a relative gain compared to baselines by over $26\%$ in F-score, and over $17\%$ in Area Under PR curve. On live search traffic, our model gains significant improvement in multiple countries.

It Takes Two to Lie: One to Lie, and One to Listen

Denis Peskov, Benny Cheng, Ahmed Elgohary, Joe Barrow, Cristian Danescu-Niculescu-Mizil and Jordan Boyd-Graber

Trust is implicit in many online text conversations---striking up new friendships, or asking for tech support. But trust can be betrayed through deception. We study the language and dynamics of deception in the negotiation-based game Diplomacy, where seven players compete for world domination by forging and breaking alliances with each other. Our study with players from the Diplomacy community gathers 17,289 messages annotated by the sender for their intended truthfulness and by the receiver for their perceived truthfulness. Unlike existing datasets, this captures deception in long-lasting relationships, where the interlocutors strategically combine truth with lies to advance objectives. A model that uses power dynamics and conversational contexts can predict when a lie occurs nearly as well as human players.

Parallel Corpus Filtering via Pre-trained Language Models

Boliang Zhang, Ajay Nagesh and Kevin Knight

Web-crawled data provides a good source of parallel corpora for training machine translation models. It is automatically obtained, but extremely noisy, and recent work shows that neural machine translation systems are more sensitive to noise than traditional statistical machine translation methods. In this paper, we propose a novel approach to filter out noisy sentence pairs from web-crawled corpora via pre-trained language models. We measure sentence parallelism by leveraging the multilingual capability of BERT and use the Generative Pre-training (GPT) language model as a domain filter to balance data domains. We evaluate the proposed method on the WMT 2018 Parallel Corpus Filtering shared task, and on our own web-crawled Japanese-Chinese parallel corpus. Our method significantly outperforms baselines and achieves a new state-of-the-art. In an unsupervised setting, our method achieves comparable performance to the top-1 supervised method. We also evaluate on a web-crawled Japanese-Chinese parallel corpus that we make publicly available.

Enhancing Pre-trained Chinese Character Representation with Word-aligned Attention

Yanzeng Li, Bowen Yu, Xue Mengge and Tingwen Liu

Most Chinese pre-trained models take character as the basic unit and learn representation according to character's external contexts, ignoring the semantics expressed in the word, which is the smallest meaningful utterance in Chinese.
Hence, we propose a novel word-aligned attention to exploit explicit word information, which is complementary to various character-based Chinese pre-trained language models.
Specifically, we devise a pooling mechanism to align the character-level attention to the word level and propose to alleviate the potential issue of segmentation error propagation by multi-source information fusion.
As a result, word and character information are explicitly integrated at the fine-tuning procedure.
Experimental results on five Chinese NLP benchmark tasks demonstrate that our method achieves significant improvements against BERT, ERNIE and BERT-wwm.

Active Learning for Coreference Resolution using Discrete Annotation

Belinda Z. Li, Gabriel Stanovsky and Luke Zettlemoyer

We improve upon pairwise annotation for active learning in coreference resolution, by asking annotators to identify mention antecedents if a presented mention pair is deemed not coreferent. This simple modification, when combined with a novel mention clustering algorithm for selecting which examples to label, is much more efficient in terms of the performance obtained per annotation budget. In experiments with existing benchmark coreference datasets, we show that the signal from this additional question leads to significant performance gains per human-annotation hour. Future work can use our annotation protocol to effectively develop coreference models for new domains. Our code is publicly available.

Demographics Should Not Be the Reason of Toxicity: Mitigating Discrimination in Text Classifications with Instance Weighting

Guanhua Zhang, Bing Bai, Junqi Zhang, Kun Bai, Conghui Zhu and Tiejun Zhao

With the recent proliferation of the use of text classifications, researchers have found that there are certain unintended biases in text classification datasets.
For example, texts containing some demographic identity-terms (e.g., "gay", "black") are more likely to be abusive in existing abusive language detection datasets.
As a result, models trained with these datasets may consider sentences like "She makes me happy to be gay" as abusive simply because of the word "gay."
In this paper, we formalize the unintended biases in text classification datasets as a kind of selection bias from the non-discrimination distribution to the discrimination distribution.
Based on this formalization, we further propose a model-agnostic debiasing training framework by recovering the non-discrimination distribution using instance weighting, which does not require any extra resources or annotations apart from a pre-defined set of demographic identity-terms.
Experiments demonstrate that our method can effectively alleviate the impacts of the unintended biases without significantly hurting models' generalization ability.

NeuInfer: Knowledge Inference on N-ary Facts

Saiping Guan, Xiaolong Jin, Jiafeng Guo, Yuanzhuo Wang and Xueqi Cheng

Knowledge inference on knowledge graph has attracted extensive attention, which aims to find out connotative valid facts in knowledge graph and is very helpful for improving the performance of many downstream applications. However, researchers have mainly poured attention to knowledge inference on binary facts. The studies on n-ary facts are relatively scarcer, although they are also ubiquitous in the real world. Therefore, this paper addresses knowledge inference on n-ary facts. We represent each n-ary fact as a primary triple coupled with a set of its auxiliary descriptive attribute-value pair(s). We further propose a neural network model, NeuInfer, for knowledge inference on n-ary facts. Besides handling the common task to infer an unknown element in a whole fact, NeuInfer can cope with a new type of task, flexible knowledge inference. It aims to infer an unknown element in a partial fact consisting of the primary triple coupled with any number of its auxiliary description(s). Experimental results demonstrate the remarkable superiority of NeuInfer.

“Who said it, and Why?” Provenance for Natural Language Claims

Yi Zhang, Zachary Ives and Dan Roth

In an era where generating content and publishing it is so easy, we are bombarded with information and are exposed to all kinds of claims, some of which do not always rank high on the truth scale.
This paper suggests that the key to a longer-term, holistic, and systematic approach to navigating this information pollution is capturing the provenance of claims.
To do that, we develop a formal definition of provenance graph for a given natural language claim, aiming to understand where the claim may come from and how it has evolved. To construct the graph, we model provenance inference, formulated mainly as an information extraction task and addressed via a textual entailment model. We evaluate our approach using two benchmark datasets, showing initial success in capturing the notion of provenance and its effectiveness on the application of claim verification.

Attend, Translate and Summarize: An Efficient Method for Neural Cross-Lingual Summarization

Junnan Zhu, Yu Zhou, Jiajun Zhang and Chengqing Zong

Cross-lingual summarization aims at summarizing a document in one language (e.g., Chinese) into another language (e.g., English). In this paper, we propose a novel method inspired by the translation pattern in the process of obtaining a cross-lingual summary. We first attend to some words in the source text, then translate them into the target language, and summarize to get the final summary. Specifically, we first employ the encoder-decoder attention distribution to attend to the source words. Second, we present three strategies to acquire the translation probability, which helps obtain the translation candidates for each source word. Finally, each summary word is generated either from the neural distribution or from the translation candidates of source words. Experimental results on Chinese-to-English and English-to-Chinese summarization tasks have shown that our proposed method can significantly outperform the baselines, achieving comparable performance with the state-of-the-art.

Improving Event Detection via Open-domain Trigger Knowledge

Meihan Tong, Bin Xu, Shuai Wang, Yixin Cao, Lei Hou, Juanzi Li and Jun Xie

Event Detection (ED) is a fundamental task in automatically structuring texts. Due to the small scale of training data, previous methods perform poorly on unseen/sparsely labeled trigger words and are prone to overfitting densely labeled trigger words. To address the issue, we propose a novel Enrichment Knowledge Distillation (EKD) model to leverage external open-domain trigger knowledge to reduce the in-built biases to frequent trigger words in annotations. Experiments on benchmark ACE2005 show that our model outperforms nine strong baselines, is especially effective for unseen/sparsely labeled trigger words. The source code is released on https://github.com/shuaiwa16/ekd.git.

Towards Interpretable Clinical Diagnosis with Bayesian Network Ensembles Stacked on Entity-Aware CNNs

Jun Chen, Xiaoya Dai, Quan Yuan, Chao Lu and Haifeng Huang

The automatic text-based diagnosis remains a challenging task for clinical use because it requires appropriate balance between accuracy and interpretability. In this paper, we attempt to propose a solution by introducing a novel framework that stacks Bayesian Network Ensembles on top of Entity-Aware Convolutional Neural Networks (CNN) towards building an accurate yet interpretable diagnosis system. The proposed framework takes advantage of the high accuracy and generality of deep neural networks as well as the interpretability of Bayesian Networks, which is critical for AI-empowered healthcare. The evaluation conducted on the real Electronic Medical Record (EMR) documents from hospitals and annotated by professional doctors proves that, the proposed framework outperforms the previous automatic diagnosis methods in accuracy performance and the diagnosis explanation of the framework is reasonable.

Why Overfitting Isn't Always Bad: Retrofitting Cross-Lingual Word Embeddings to Dictionaries

Mozhi Zhang, Yoshinari Fujinuma, Michael J. Paul and Jordan Boyd-Graber

Cross-lingual word embeddings (CLWE) are often evaluated on bilingual lexicon induction (BLI). Recent CLWE methods use linear projections, which underfit the training dictionary, to generalize on BLI. However, underfitting can hinder generalization to other downstream tasks that rely on words from the training dictionary. We address this limitation by retrofitting CLWE to the training dictionary, which pulls training translation pairs closer in the embedding space and overfits the training dictionary. This simple post-processing step often improves accuracy on two downstream tasks, despite lowering BLI test accuracy. We also retrofit to both the training dictionary and a synthetic dictionary induced from CLWE, which sometimes generalizes even better on downstream tasks. Our results confirm the importance of fully exploiting training dictionary in downstream tasks and explains why BLI is a flawed CLWE evaluation.

Multi-Domain Dialogue Acts and Response Co-Generation

Kai Wang, Junfeng Tian, Rui Wang, Xiaojun Quan and Jianxing Yu

Generating fluent and informative responses is of critical importance for task-oriented dialogue systems. Existing pipeline approaches generally predict multiple dialogue acts first and use them to assist response generation. There are at least two shortcomings with such approaches. First, the inherent structures of multi-domain dialogue acts are neglected. Second, the semantic associations between acts and responses are not taken into account for response generation. To address these issues, we propose a neural co-generation model that generates dialogue acts and responses concurrently. Unlike those pipeline approaches, our act generation module preserves the semantic structures of multi-domain dialogue acts and our response generation module dynamically attends to different acts as needed. We train the two modules jointly using an uncertainty loss to adjust their task weights adaptively. Extensive experiments are conducted on the large-scale MultiWOZ dataset and the results show that our model achieves very favorable improvement over several state-of-the-art models in both automatic and human evaluations.

Efficient Pairwise Annotation of Argument Quality

Lukas Gienapp, Benno Stein, Matthias Hagen and Martin Potthast

We present an efficient annotation framework for argument quality, a feature difficult to be measured reliably as per previous work. A stochastic transitivity model is combined with an effective sampling strategy to infer high-quality labels with low effort from crowdsourced pairwise judgments. The model's capabilities are showcased by compiling Webis-ArgQuality-20, an argument quality corpus that comprises scores for rhetorical, logical, dialectical, and overall quality inferred from a total of 41,859 pairwise judgments among 1,271 arguments. With up to 93% cost savings, our approach significantly outperforms existing annotation procedures. Furthermore, novel insight into argument quality is provided through statistical analysis, and a new aggregation method to infer overall quality from individual quality dimensions is proposed.

Adversarial NLI: A New Benchmark for Natural Language Understanding

Yixin Nie, Adina Williams, Emily Dinan, Mohit Bansal, Jason Weston and Douwe Kiela

We introduce a new large-scale NLI benchmark dataset, collected via an iterative, adversarial human-and-model-in-the-loop procedure. We show that training models on this new dataset leads to state-of-the-art performance on a variety of popular NLI benchmarks, while posing a more difficult challenge with its new test set. Our analysis sheds light on the shortcomings of current state-of-the-art models, and shows that non-expert annotators are successful at finding their weaknesses. The data collection method can be applied in a never-ending learning scenario, becoming a moving target for NLU, rather than a static benchmark that will quickly saturate.

ParaCrawl: Web-Scale Acquisition of Parallel Corpora

Marta Bañón, Pinzhen Chen, Barry Haddow, Kenneth Heafield, Hieu Hoang, Miquel Esplà-Gomis, Mikel L. Forcada, Amir Kamran, Faheem Kirefu, Philipp Koehn, Sergio Ortiz Rojas, Leopoldo Pla Sempere, Gema Ramírez-Sánchez, Elsa Sarrías, Marek Strelec, Brian Thompson, William Waites, Dion Wiggins and Jaume Zaragoza

We report on methods to create the largest publicly available parallel corpora by crawling the web, using open source software. We empirically compare alternative methods and publish benchmark data sets for sentence alignment and sentence pair filtering. We also describe the parallel corpora released and evaluate their quality and their usefulness to create machine translation systems.

tBERT: Topic Models and BERT Joining Forces for Semantic Similarity Detection

Nicole Peinelt, Dong Nguyen and Maria Liakata

Semantic similarity detection is a fundamental task in natural language understanding. Adding topic information has been useful for previous feature-engineered semantic similarity models as well as neural models for other tasks. There is currently no standard way of combining topics with pretrained contextual representations such as BERT. We propose a novel topic-informed BERT-based architecture for pairwise semantic similarity detection and show that our model improves performance over strong neural baselines across a variety of English language datasets. We find that the addition of topics to BERT helps particularly with resolving domain-specific cases.

Multimodal and Multiresolution Speech Recognition with Transformers

Georgios Paraskevopoulos, Srinivas Parthasarathy, Aparna Khare and Shiva Sundaram

This paper presents an audio visual automatic speech recognition (AV-ASR) system using a Transformer-based architecture. We particularly focus on the scene context provided by the visual information, to ground the ASR. We extract representations for audio features in the encoder layers of the transformer and fuse video features using an additional crossmodal multihead attention layer. Additionally, we incorporate a multitask training criterion for multiresolution ASR, where we train the model to generate both character and subword level transcriptions.
Experimental results on the How2 dataset, indicate that multiresolution training can speed up convergence by around 50\% and relatively improves word error rate (WER) performance by upto 18% over subword prediction models. Further, incorporating visual information improves performance with relative gains upto 3.76% over audio only models.
Our results are comparable to state-of-the-art Listen, Attend and Spell-based architectures.

Keyphrase Generation for Scientific Document Retrieval

Florian Boudin, Ygor Gallina and Akiko Aizawa

Sequence-to-sequence models have lead to significant progress in keyphrase generation, but it remains unknown whether they are reliable enough to be beneficial for document retrieval. This study provides empirical evidence that such models can significantly improve retrieval performance, and introduces a new extrinsic evaluation framework that allows for a better understanding of the limitations of keyphrase generation models. Using this framework, we point out and discuss the difficulties encountered with supplementing documents with -not present in text- keyphrases, and generalizing models across domains. Our code is available at https://github.com/boudinfl/ir-using-kg

BERTRAM: Improved Word Embeddings Have Big Impact on Contextualized Model Performance

Timo Schick and Hinrich Schütze

Pretraining deep language models has led to large performance gains in NLP. Despite this success, Schick and Schütze (2020) recently showed that these models struggle to understand rare words.
For static word embeddings, this problem has been addressed by separately learning representations for rare words. In this work, we transfer this idea to pretrained language models: We introduce BERTRAM, a powerful architecture based on BERT that is capable of inferring high-quality embeddings for rare words that are suitable as input representations for deep language models. This is achieved by enabling the surface form and contexts of a word to interact with
each other in a deep architecture.
Integrating BERTRAM into BERT leads to large performance increases due to improved representations of rare and medium frequency words on both a rare word probing task and three downstream tasks.

DTCA: Decision Tree-based Co-Attention Networks for Explainable Claim Verification

Lianwei Wu, Yuan Rao, Yongqiang Zhao, Hao Liang and Ambreen Nazir

Recently, many methods discover effective evidence from reliable sources by appropriate neural networks for explainable claim verification, which has been widely recognized. However, in these methods, the discovery process of evidence is nontransparent and unexplained. Simultaneously, the discovered evidence is aimed at the interpretability of the whole sequence of claims but insufficient to focus on the false parts of claims. In this paper, we propose a Decision Tree-based Co-Attention model (DTCA) to discover evidence for explainable claim verification. Specifically, we first construct Decision Tree-based Evidence model (DTE) to select comments with high credibility as evidence in a transparent and interpretable way. Then we design Co-attention Self-attention networks (CaSa) to make the selected evidence interact with claims, which is for 1) training DTE to determine the optimal decision thresholds and obtain more powerful evidence; and 2) utilizing the evidence to find the false parts in the claim. Experiments on two public datasets, RumourEval and PHEME, demonstrate that DTCA not only provides explanations for the results of claim verification but also achieves the state-of-the-art performance, boosting the F1-score by more than 3.11%, 2.41%, respectively.

Meta-Transfer Learning for Code-Switched Speech Recognition

Genta Indra Winata, Samuel Cahyawijaya, Zhaojiang Lin, Zihan Liu, Peng Xu and Pascale Fung

An increasing number of people in the world today speak a mixed-language as a result of being multilingual. However, building a speech recognition system for code-switching remains difficult due to the availability of limited resources and the expense and significant effort required to collect mixed-language data. We therefore propose a new learning method, meta-transfer learning, to transfer learn on a code-switched speech recognition system in a low-resource setting by judiciously extracting information from high-resource monolingual datasets. Our model learns to recognize individual languages, and transfer them so as to better recognize mixed-language speech by conditioning the optimization on the code-switching data. Based on experimental results, our model outperforms existing baselines on speech recognition and language modeling tasks, and is faster to converge.

Structural Information Preserving for Graph-to-Text Generation

Linfeng Song, Ante Wang, Jinsong Su, Yue Zhang, Kun Xu, Yubin Ge and Dong Yu

The task of graph-to-text generation aims at producing sentences that preserve the meaning of input graphs. As a crucial defect, the current state-of-the-art models may mess up or even drop the core structural information of input graphs when generating outputs. We propose to tackle this problem by leveraging richer training signals that can guide our model for preserving input information. In particular, we introduce two types of autoencoding losses, each individually focusing on different aspects (a.k.a. views) of input graphs. The losses are then back-propagated to better calibrate our model via multi-task training. Experiments on two benchmarks for graph-to-text generation show the effectiveness of our approach over a state-of-the-art baseline.

ZPR2: Joint Zero Pronoun Recovery and Resolution using Multi-Task Learning and BERT

Linfeng Song, Kun Xu, Yue Zhang, Jianshu Chen and Dong Yu

Zero pronoun recovery and resolution aim at recovering the dropped pronoun and pointing out its anaphoric mentions, respectively. We propose to better explore their interaction by solving both tasks together, while the previous work treats them separately. For zero pronoun resolution, we study this task in a more realistic setting, where no parsing trees or only automatic trees are available, while most previous work assumes gold trees. Experiments on two benchmarks show that joint modeling significantly outperforms our baseline that already beats the previous state of the arts.

DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference

Ji Xin, Raphael Tang, Jaejun Lee, Yaoliang Yu and Jimmy Lin

Large-scale pre-trained language models such as BERT have brought significant improvements to NLP applications. However, they are also notorious for being slow in inference, which makes them difficult to deploy in real-time applications. We propose a simple but effective method, DeeBERT, to accelerate BERT inference. Our approach allows samples to exit earlier without passing through the entire model. Experiments show that DeeBERT is able to save up to ~40% inference time with minimal degradation in model quality. Further analyses show different behaviors in the BERT transformer layers and also reveal their redundancy. Our work provides new ideas to efficiently apply deep transformer-based models to downstream tasks. Code is available at https://github.com/castorini/DeeBERT.

On Exposure Bias, Hallucination and Domain Shift in Neural Machine Translation

Chaojun Wang and Rico Sennrich

The standard training algorithm in neural machine translation (NMT) suffers from exposure bias, and alternative algorithms have been proposed to mitigate this. However, the practical impact of exposure bias is under debate. In this paper, we link exposure bias to another well-known problem in NMT, namely the tendency to generate hallucinations under domain shift. In experiments on three datasets with multiple test domains, we show that exposure bias is partially to blame for hallucinations, and that training with Minimum Risk Training, which avoids exposure bias, can mitigate this. Our analysis explains why exposure bias is more problematic under domain shift, and also links exposure bias to the beam search problem, i.e. performance deterioration with increasing beam size. Our results provide a new justification for methods that reduce exposure bias: even if they do not increase performance on in-domain test sets, they can increase model robustness to domain shift.

Moving Down the Long Tail of Word Sense Disambiguation with Gloss Informed Bi-encoders

Terra Blevins and Luke Zettlemoyer

A major obstacle in Word Sense Disambiguation (WSD) is that word senses are not uniformly distributed, causing existing models to generally perform poorly on senses that are either rare or unseen during training. We propose a bi-encoder model that independently embeds (1) the target word with its surrounding context and (2) the dictionary definition, or gloss, of each sense. The encoders are jointly optimized in the same representation space, so that sense disambiguation can be performed by finding the nearest sense embedding for each target word embedding. Our system outperforms previous state-of-the-art models on English all-words WSD; these gains predominantly come from improved performance on rare senses, leading to a 31.1% error reduction on less frequent senses over prior work. This demonstrates that rare senses can be more effectively disambiguated by modeling their definitions.

Document-Level Event Role Filler Extraction using Multi-Granularity Contextualized Encoding

Xinya Du and Claire Cardie

Few works in the literature of event extraction have gone beyond individual sentences to make extraction decisions. This is problematic when the information needed to recognize an event argument is spread across multiple sentences. We argue that document-level event extraction is a difficult task since it requires a view of a larger context to determine which spans of text correspond to event role fillers. We first investigate how end-to-end neural sequence models (with pre-trained language model representations) perform on document-level role filler extraction, as well as how the length of context captured affects the models’ performance. To dynamically aggregate information captured by neural representations learned at different levels of granularity (e.g., the sentence- and paragraph-level), we propose a novel multi-granularity reader. We evaluate our models on the MUC-4 event extraction dataset, and show that our best system performs substantially better than prior work. We also report findings on the relationship between context length and neural model performance on the task.

A Top-down Neural Architecture towards Text-level Parsing of Discourse Rhetorical Structure

Longyin Zhang, Yuqing Xing, Fang Kong, Peifeng Li and Guodong Zhou

Due to its great importance in deep natural language understanding and various down-stream applications, text-level parsing of discourse rhetorical structure (DRS) has been drawing more and more attention in recent years.
However, all the previous studies on text-level discourse parsing adopt bottom-up approaches, which much limit the DRS determination on local information and fail to well benefit from global information of the overall discourse.
In this paper, we justify from both computational and perceptive points-of-view that the top-down architecture is more suitable for text-level DRS parsing. On the basis, we propose a top-down neural architecture toward text-level DRS parsing.
In particular, we cast discourse parsing as a recursive split point ranking task, where a split point is classified to different levels according to its rank and the elementary discourse units (EDUs) associated with it are arranged accordingly. In this way, we can determine the complete DRS as a hierarchical tree structure via an encoder-decoder with an internal stack. Experimentation on both the English RST-DT corpus and the Chinese CDTB corpus shows the great effectiveness of our proposed top-down approach towards text-level DRS parsing.

Enriched In-Order Linearization for Faster Sequence-to-Sequence Constituent Parsing

Daniel Fernández-González and Carlos Gómez-Rodríguez

Sequence-to-sequence constituent parsing requires a linearization to represent trees as sequences. Top-down tree linearizations, which can be based on brackets or shift-reduce actions, have achieved the best accuracy to date. In this paper, we show that these results can be improved by using an in-order linearization instead. Based on this observation, we implement an enriched in-order shift-reduce linearization inspired by Vinyals et al. (2015)'s approach, achieving the best accuracy to date on the English PTB dataset among fully-supervised single-model sequence-to-sequence constituent parsers.
Finally, we apply deterministic attention mechanisms to match the speed of state-of-the-art transition-based parsers, thus showing that sequence-to-sequence models can match them, not only in
accuracy, but also in speed.

FastBERT: a Self-distilling BERT with Adaptive Inference Time

Weijie Liu, Peng Zhou, Zhiruo Wang, Zhe Zhao, Haotang Deng and QI JU

Pre-trained language models like BERT have proven to be highly performant. However, they are often computationally expensive in many practical scenarios, for such heavy models can hardly be readily implemented with limited resources. To improve their efficiency with an assured model performance, we propose a novel speed-tunable FastBERT with adaptive inference time. The speed at inference can be flexibly adjusted under varying demands, while redundant calculation of samples is avoided. Moreover, this model adopts a unique self-distillation mechanism at fine-tuning, further enabling a greater computational efficacy with minimal loss in performance. Our model achieves promising results in twelve English and Chinese datasets. It is able to speed up by a wide range from 1 to 12 times than BERT if given different speedup thresholds to make a speed-performance tradeoff.

Diverse and Informative Dialogue Generation with Context-Specific Commonsense Knowledge Awareness

Sixing Wu, Ying Li, Dawei Zhang, Yang Zhou and Zhonghai Wu

Generative dialogue systems tend to produce generic responses, which often leads to boring conversations. For alleviating this issue, Recent studies proposed to retrieve and introduce knowledge facts from knowledge graphs. While this paradigm works to a certain extent, it usually retrieves knowledge facts only based on the entity word itself, without considering the specific dialogue context. Thus, the introduction of the context-irrelevant knowledge facts can impact the quality of generations. To this end, this paper proposes a novel commonsense knowledge-aware dialogue generation model, ConKADI. We design a Felicitous Fact mechanism to help the model focus on the knowledge facts that are highly relevant to the context; furthermore, two techniques, Context-Knowledge Fusion and Flexible Mode Fusion are proposed to facilitate the integration of the knowledge in the ConKADI. We collect and build a large-scale Chinese dataset aligned with the commonsense knowledge for dialogue generation. Extensive evaluations over both an open-released English dataset and our Chinese dataset demonstrate that our approach ConKADI outperforms the state-of-the-art approach CCM, in most experiments.

Unsupervised Opinion Summarization with Noising and Denoising

Reinald Kim Amplayo and Mirella Lapata

The supervised training of high-capacity models on large datasets containing hundreds of thousands of document-summary pairs is critical to the recent success of deep learning techniques for abstractive summarization. Unfortunately, in most domains (other than news) such training data is not available and cannot be easily sourced. In this paper we enable the use of supervised learning for the setting where there are only documents available (e.g., product or business reviews) without ground truth summaries. We create a synthetic dataset from a corpus of user reviews by sampling a review, pretending it is a summary, and generating noisy versions thereof which we treat as pseudo-review input. We introduce several linguistically motivated noise generation functions and a summarization model which learns to denoise the input and generate the original review. At test time, the model accepts genuine reviews and generates a summary containing salient opinions, treating those that do not reach consensus as noise. Extensive automatic and human evaluation shows that our model brings substantial improvements over both abstractive and extractive baselines.

Tchebycheff Procedure for Multi-task Text Classification

Yuren Mao, Shuang Yun, Weiwei Liu and Bo Du

Multi-task Learning methods have achieved great progress in text classification. However, existing methods assume that multi-task text classification problems are convex multiobjective optimization problems, which is unrealistic in real-world applications. To address this issue, this paper presents a novel Tchebycheff procedure to optimize the multi-task classification problems without convex assumption. The extensive experiments back up our theoretical analysis and validate the superiority of our proposals.

A Span-based Linearization for Constituent Trees

Yang Wei, Yuanbin Wu and Man Lan

We propose a novel linearization of a constituent tree, together with a new locally normalized model. For each split point in a sentence, our model computes the normalizer on all spans ending with that split point, and then predicts a tree span from them. Compared with global models, our model is fast and parallelizable. Different from previous local models, our linearization method is tied on the spans directly and considers more local features when performing span prediction, which is more interpretable and effective. Experiments on PTB (95.8 F1) and CTB (92.4 F1) show that our model significantly outperforms existing local models and efficiently achieves competitive results with global models.

Curriculum Pre-training for End-to-End Speech Translation

Chengyi Wang, Yu Wu, Shujie Liu, Ming Zhou and Zhenglu Yang

End-to-end speech translation poses a heavy burden on the encoder because it has to transcribe, understand, and learn cross-lingual semantics simultaneously. To obtain a powerful encoder, traditional methods pre-train it on ASR data to capture speech features. However, we argue that pre-training the encoder only through simple speech recognition is not enough, and high-level linguistic knowledge should be considered. Inspired by this, we propose a curriculum pre-training method that includes an elementary course for transcription learning and two advanced courses for understanding the utterance and mapping words in two languages. The difficulty of these courses is gradually increasing. Experiments show that our curriculum pre-training method leads to significant improvements on En-De and En-Fr speech translation benchmarks.

Self-Attention with Cross-Lingual Position Representation

Liang Ding, Longyue Wang and Dacheng Tao

Position encoding (PE), an essential part of self-attention networks (SANs), is used to preserve the word order information for natural language processing tasks, generating fixed position indices for input sequences. However, in cross-lingual scenarios, \eg machine translation, the PEs of source and target sentences are modeled independently. Due to word order divergences in different languages, modeling the cross-lingual positional relationships might help SANs tackle this problem.
In this paper, we augment SANs with \emph{cross-lingual position representations} to model the bilingually aware latent structure for the input sentence. Specifically, we utilize bracketing transduction grammar (BTG)-based reordering information to encourage SANs to learn bilingual diagonal alignments. Experimental results on WMT'14 English$\Rightarrow$German, WAT'17 Japanese$\Rightarrow$English, and WMT'17 Chinese$\Leftrightarrow$English translation tasks demonstrate that our approach significantly and consistently improves translation quality over strong baselines. Extensive analyses confirm that the performance gains come from the cross-lingual information.

Reducing Gender Bias in Neural Machine Translation as a Domain Adaptation Problem

Danielle Saunders and Bill Byrne

Training data for NLP tasks often exhibits gender bias in that fewer sentences refer to women than to men. In Neural Machine Translation (NMT) gender bias has been shown to reduce translation quality, particularly when the target language has grammatical gender. The recent WinoMT challenge set allows us to measure this effect directly (Stanovsky et al, 2019)

Ideally we would reduce system bias by simply debiasing all data prior to training, but achieving this effectively is itself a challenge. Rather than attempt to create a `balanced' dataset, we use transfer learning on a small set of trusted, gender-balanced examples. This approach gives strong and consistent improvements in gender debiasing with much less computational cost than training from scratch.

A known pitfall of transfer learning on new domains is `catastrophic forgetting', which we address at adaptation and inference time. During adaptation we show that Elastic Weight Consolidation allows a performance trade-off between general translation quality and bias reduction. At inference time we propose a lattice-rescoring scheme which outperforms all systems evaluated in Stanovsky et al, 2019 on WinoMT with no degradation of general test set BLEU. We demonstrate our approach translating from English into three languages with varied linguistic properties and data availability.

Efficient Second-Order TreeCRF for Neural Dependency Parsing

Yu Zhang, Zhenghua Li and Min Zhang

In the deep learning (DL) era, parsing models are extremely simplified with little hurt on performance, thanks to the remarkable capability of multi-layer BiLSTMs in context representation.
As the most popular graph-based dependency parser due to its high efficiency and performance, the biaffine parser directly scores single dependencies under the arc-factorization assumption, and adopts a very simple local token-wise cross-entropy training loss.
This paper for the first time presents a second-order TreeCRF extension to the biaffine parser.
For a long time, the complexity and inefficiency of the inside-outside algorithm hinder the popularity of TreeCRF.
To address this issue, we propose an effective way to batchify the inside and Viterbi algorithms for direct large matrix operation on GPUs, and to avoid the complex outside algorithm via efficient back-propagation.
Experiments and analysis on 27 datasets from 13 languages clearly show that techniques developed before the DL era, such as structural learning (global TreeCRF loss) and high-order modeling are still useful, and can further boost parsing performance over the state-of-the-art biaffine parser, especially for partially annotated training data.
We release our code at https://github.com/yzhangcs/crfpar.

SAFER: A Structure-free Approach for Certified Robustness to Adversarial Word Substitutions

Mao Ye, Chengyue Gong and Qiang Liu

State-of-the-art NLP models can often be fooled by human-unaware transformations such as synonymous word substitution. For security reasons, it is of critical importance to develop models with certified robustness that can provably guarantee that the prediction is can not be altered by any possible synonymous word substitution. In this work, we propose a certified robust method based on a new randomized smoothing technique, which constructs a stochastic ensemble by applying random word substitutions on the input sentences, and leverage the statistical properties of the ensemble to provably certify the robustness. Our method is simple and structure-free in that it only requires the black-box queries of the model outputs, and hence can be applied to any pre-trained models (such as BERT) and any types of models (world-level or subword-level). Our method significantly outperforms recent state-of-the-art methods for certified robustness on both IMDB and Amazon text classification tasks. To the best of our knowledge, we are the first work to achieve certified robustness on large systems such as BERT with practically meaningful certified accuracy.

Storytelling with Dialogue: {A} {Critical Role Dungeons and Dragons Dataset}

Revanth Rameshkumar and Peter Bailey

This paper describes the {Critical Role Dungeons and Dragons Dataset} ({CRD3}) and related analyses. {Critical Role} is an unscripted, live-streamed show where a fixed group of people play {Dungeons and Dragons}, an open-ended role-playing game. The dataset is collected from 159 {Critical Role} episodes transcribed to text dialogues, consisting of 398,682 turns. It also includes corresponding abstractive summaries collected from the {Fandom} wiki. The dataset is linguistically unique in that the narratives are generated entirely through player collaboration and spoken interaction. For each dialogue, there are a large number of turns, multiple abstractive summaries with varying levels of detail, and semantic ties to the previous dialogues. {I}n addition, we provide a data augmentation method that produces 34,243 summary-dialogue chunk pairs to support current neural {ML} approaches, and we provide an abstractive summarization benchmark and evaluation.

Frugal Paradigm Completion

Alexander Erdmann, Tom Kenter, Markus Becker and Christian Schallhart

Lexica distinguishing all morphologically related forms of each lexeme are crucial to many language technologies, yet building them is expensive. We propose a frugal paradigm completion approach that predicts all related forms in a morphological paradigm from as few manually provided forms as possible. It induces typological information during training which it uses to determine the best sources at test time. We evaluate our language-agnostic approach on 7 diverse languages. Compared to popular alternative approaches, ours reduces manual labor by 16-63% and is the most robust to typological variation.

Improved Speech Representations with Multi-Target Autoregressive Predictive Coding

Yu-An Chung and James Glass

Training objectives based on predictive coding have recently been shown to be very effective at learning meaningful representations from unlabeled speech. One example is Autoregressive Predictive Coding (Chung et al., 2019), which trains an autoregressive RNN to generate an unseen future frame given a context such as recent past frames. The basic hypothesis of these approaches is that hidden states that can accurately predict future frames are a useful representation for many downstream tasks. In this paper we extend this hypothesis and aim to enrich the information encoded in the hidden states by training the model to make more accurate future predictions. We propose an auxiliary objective that serves as a regularization to improve generalization of the future frame prediction task. Experimental results on phonetic classification, speech recognition, and speech translation not only support the hypothesis, but also demonstrate the effectiveness of our approach in learning representations that contain richer phonetic content.

Low-Dimensional Hyperbolic Knowledge Graph Embeddings

Ines Chami, Adva Wolf, Da-Cheng Juan, Frederic Sala, Sujith Ravi and Christopher Ré

Knowledge graph (KG) embeddings learn low- dimensional representations of entities and relations to predict missing facts. KGs often exhibit hierarchical and logical patterns which must be preserved in the embedding space. For hierarchical data, hyperbolic embedding methods have shown promise for high-fidelity and parsimonious representations. However, existing hyperbolic embedding methods do not account for the rich logical patterns in KGs. In this work, we introduce a class of hyperbolic KG embedding models that simultaneously capture hierarchical and logical patterns. Our approach combines hyperbolic reflections and rotations with attention to model complex relational patterns. Experimental results on standard KG benchmarks show that our method improves over previous Euclidean- and hyperbolic-based efforts by up to 6.1% in mean reciprocal rank (MRR) in low dimensions. Furthermore, we observe that different geometric transformations capture different types of relations while attention- based transformations generalize to multiple relations. In high dimensions, our approach yields new state-of-the-art MRRs of 49.6% on WN18RR and 57.7% on YAGO3-10.

Learning Interpretable Relationships between Entities, Relations and Concepts via Bayesian Structure Learning on Open Domain Facts

Jingyuan Zhang, Mingming Sun, Yue Feng and Ping Li

Concept graphs are created as universal taxonomies for text understanding in the open-domain knowledge. The nodes in concept graphs include both entities and concepts. The edges are from entities to concepts, showing that an entity is an instance of a concept. In this paper, we propose the task of learning interpretable relationships from open-domain facts to enrich and refine concept graphs. The Bayesian network structures are learned from open-domain facts as the interpretable relationships between relations of facts and concepts of entities. We conduct extensive experiments on public English and Chinese datasets. Compared to the state-of-the-art methods, the learned network structures help improving the identification of concepts for entities based on the relations of entities on both datasets.

SPECTER: Document-level Representation Learning using Citation-informed Transformers

Arman Cohan, Sergey Feldman, Iz Beltagy, Doug Downey and Daniel Weld

Representation learning is a critical ingredient for natural language processing systems. Recent Transformer language models like BERT learn powerful textual representations, but these models are targeted towards token- and sentence-level training objectives and do not leverage information on inter-document relatedness, which limits their document-level representation power.
For applications on scientific documents, such as classification and recommendation, accurate embeddings of documents are a necessity. We propose SPECTER, a new method to generate document-level embedding of scientific papers based on pretraining a Transformer language model on a powerful signal of document-level relatedness: the citation graph. Unlike existing pretrained language models, Specter can be easily applied to downstream applications without task-specific fine-tuning.
Additionally, to encourage further research on document-level models, we introduce SciDocs, a new evaluation benchmark consisting of seven document-level tasks ranging from citation prediction, to document classification and recommendation. We show that Specter outperforms a variety of competitive baselines on the benchmark.

On the Encoder-Decoder Incompatibility in Variational Text Modeling and Beyond

Chen Wu, Prince Zizhuang Wang and William Yang Wang

Variational autoencoders (VAEs) combine latent variables with amortized variational inference, whose optimization usually converges into a trivial local optimum termed posterior collapse, especially in text modeling. By tracking the optimization dynamics, we observe the encoder-decoder incompatibility that leads to poor parameterizations of the data manifold. We argue that the trivial local optimum may be avoided by improving the encoder and decoder parameterizations since the posterior network is part of a transition map between them. To this end, we propose Coupled-VAE, which couples a VAE model with a deterministic autoencoder with the same structure and improves the encoder and decoder parameterizations via encoder weight sharing and decoder signal matching. We apply the proposed Coupled-VAE approach to various VAE models with different regularization, posterior family, decoder structure, and optimization strategy. Experiments on benchmark datasets (i.e., PTB, Yelp, and Yahoo) show consistently improved results in terms of probability estimation and richness of the latent space. We also generalize our method to conditional language modeling and propose Coupled-CVAE, which largely improves the diversity of dialogue generation on the Switchboard dataset.

Translationese as a Language in "Multilingual" NMT

Parker Riley, Isaac Caswell, Markus Freitag and David Grangier

Machine translation has an undesirable propensity to produce ``translationese" artifacts, which can lead to higher BLEU scores while being liked less by human raters. Motivated by this, we model translationese and original (i.e. natural) text as separate languages in a multilingual model, and pose the question: can we perform zero-shot translation between original source text and original target text? There is no data with original source and original target, so we train a sentence-level classifier to distinguish translationese from original target text, and use this classifier to tag the training data for an NMT model. Using this technique we bias the model to produce more natural outputs at test time, yielding gains in human evaluation scores on both accuracy and fluency. Additionally, we demonstrate that it is possible to bias the model to produce translationese and game the BLEU score, increasing it while decreasing human-rated quality. We analyze these outputs using metrics measuring the degree of translationese, and present an analysis of the volatility of heuristic-based train-data tagging.

Taxonomy Construction of Unseen Domains via Graph-based Cross-Domain Knowledge Transfer

Chao Shang, Sarthak Dash, Md. Faisal Mahbub Chowdhury, Nandana Mihindukulasooriya and Alfio Gliozzo

Extracting lexico-semantic relations as graph-structured taxonomies, also known as taxonomy construction, has been beneficial in a variety of NLP applications. Recently Graph Neural Network (GNN) has shown to be powerful in successfully tackling many tasks. However, there has been no attempt to exploit GNN to create taxonomies. In this paper, we propose Graph2Taxo, a GNN-based cross-domain transfer framework for the taxonomy construction task. Our main contribution is to learn the latent features of taxonomy construction from existing domains to guide the structure learning of an unseen domain. We also propose a novel method of directed acyclic graph (DAG) generation for taxonomy construction. Specifically, our proposed Graph2Taxo uses a noisy graph constructed from automatically extracted noisy hyponym hypernym candidate pairs, and a set of taxonomies for some known domains for training. The learned model is then used to generate taxonomy for a new unknown domain given a set of terms for that domain. Experiments on benchmark datasets from science and environment domains show that our approach attains significant improvements correspondingly over the state of the art.

Improving Entity Linking through Semantic Reinforced Entity Embeddings

Feng Hou, Ruili Wang, Jun He and Yi Zhou

Entity embeddings, which represent different aspects of each entity with a single vector like word embeddings, are a key component of neural entity linking models. Existing entity embeddings are learned from canonical Wikipedia articles and local contexts surrounding target entities. Such entity embeddings are effective, but too distinctive for linking models to learn contextual commonality. We propose a simple yet effective method, FGS2EE, to inject fine-grained semantic information into entity embeddings to reduce the distinctiveness and facilitate the learning of contextual commonality. FGS2EE first uses the embeddings of semantic type words to generate semantic embeddings, and then combines them with existing entity embeddings through linear aggregation. Extensive experiments show the effectiveness of such embeddings. Based on our entity embeddings, we achieved new sate-of-the-art performance on entity linking.

Unsupervised Paraphrasing by Simulated Annealing

Xianggen Liu, Lili Mou, Fandong Meng, Hao Zhou, Jie Zhou and Sen Song

We propose UPSA, a novel approach that accomplishes Unsupervised Paraphrasing by Simulated Annealing. We model paraphrase generation as an optimization problem and propose a sophisticated objective function, involving semantic similarity, expression diversity, and language fluency of paraphrases. UPSA searches the sentence space towards this objective by performing a sequence of local editing. We evaluate our approach on various datasets, namely, Quora, Wikianswers, MSCOCO, and Twitter. Extensive results show that UPSA achieves the state-of-the-art performance compared with previous unsupervised methods in terms of both automatic and human evaluations. Further, our approach outperforms most existing domain-adapted supervised models, showing the generalizability of UPSA.

MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices

Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang and Denny Zhou

Natural Language Processing (NLP) has recently achieved great success by using huge pre-trained models with hundreds of millions of parameters. However, these models suffer from heavy model sizes and high latency such that they cannot be deployed to resource-limited mobile devices. In this paper, we propose MobileBERT for compressing and accelerating the popular BERT model. Like the original BERT, MobileBERT is task-agnostic, that is, it can be generically applied to various downstream NLP tasks via simple fine-tuning. Basically, MobileBERT is a thin version of BERT_LARGE, while equipped with bottleneck structures and a carefully designed balance between self-attentions and feed-forward networks. To train MobileBERT, we first train a specially designed teacher model, an inverted-bottleneck incorporated BERT_LARGE model. Then, we conduct knowledge transfer from this teacher to MobileBERT. Empirical studies show that MobileBERT is 4.3x smaller and 5.5x faster than BERT_BASE while achieving competitive results on well-known benchmarks. On the natural language inference tasks of GLUE, MobileBERT achieves a GLUE score of 77.7 (0.6 lower than BERT_BASE), and 62 ms latency on a Pixel 4 phone. On the SQuAD v1.1/v2.0 question answering task, MobileBERT achieves a dev F1 score of 90.0/79.2 (1.5/2.1 higher than BERT_BASE).

Discourse as a Function of Event: Profiling Discourse Structure in News Articles around the Main Event

Prafulla Kumar Choubey, Aaron Lee, Ruihong Huang and Lu Wang

Understanding discourse structures of news articles is vital to effectively contextualize the occurrence of a news event. To enable computational modeling of news structures, we apply an existing theory of functional discourse structure for news articles that revolves around the main event and create a human-annotated corpus of 802 documents spanning over four domains and three media sources. Next, we propose several document-level neural-network models to automatically construct news content structures. Finally, we demonstrate that incorporating system predicted news structures yields new state-of-the-art performance for event coreference resolution. The news documents we annotated are openly available and the annotations are publicly released for future research.

Coach: A Coarse-to-Fine Approach for Cross-domain Slot Filling

Zihan Liu, Genta Indra Winata, Peng Xu and Pascale Fung

As an essential task in task-oriented dialog systems, slot filling requires extensive training data in a certain domain. However, such data are not always available. Hence, cross-domain slot filling has naturally arisen to cope with this data scarcity problem. In this paper, we propose a Coarse-to-fine approach (Coach) for cross-domain slot filling. Our model first learns the general pattern of slot entities by detecting whether the tokens are slot entities or not. It then predicts the specific types for the slot entities. In addition, we propose a template regularization approach to improve the adaptation robustness by regularizing the representation of utterances based on utterance templates. Experimental results show that our model significantly outperforms state-of-the-art approaches in slot filling. Furthermore, our model can also be applied to the cross-domain named entity recognition task, and it achieves better adaptation performance than other existing baselines. The code is available at https://github.com/zliucr/coach.

{C}onversational {W}ord {E}mbedding for {R}etrieval-{B}ased {D}ialog {S}ystem

Wentao Ma, Yiming Cui, Ting Liu, Dong Wang, Shijin Wang and Guoping Hu

Human conversations contain many types of information, e.g., knowledge, common sense, and language habits. In this paper, we propose a conversational word embedding method named PR-Embedding, which utilizes the conversation pairs <post, reply> to learn word embedding. Different from previous works, PR-Embedding uses the vectors from two different semantic spaces to represent the words in post and reply.To catch the information among the pair, we first introduce the word alignment model from statistical machine translation to generate the cross-sentence window, then train the embedding on word-level and sentence-level.We evaluate the method on single-turn and multi-turn response selection tasks for retrieval-based dialog systems.The experiment results show that PR-Embedding can improve the quality of the selected response.

Lexically Constrained Neural Machine Translation with Levenshtein Transformer

Raymond Hendy Susanto, Shamil Chollampatt and Liling Tan

This paper proposes a simple and effective algorithm for incorporating lexical constraints in neural machine translation. Previous work either required re-training existing models with the lexical constraints or incorporating them during beam search decoding with significantly higher computational overheads. Leveraging the flexibility and speed of a recently proposed Levenshtein Transformer model (Gu et al., 2019), our method injects terminology constraints at inference time without any impact on decoding speed. Our method does not require any modification to the training procedure and can be easily applied at runtime with custom dictionaries. Experiments on English-German WMT datasets show that our approach improves an unconstrained baseline and previous approaches.

From SPMRL to NMRL: What Did We Learn (and Unlearn) in a Decade of Parsing Morphologically-Rich Languages (MRLs)?

Reut Tsarfaty, Dan Bareket, Stav Klein and Amit Seker

It has been exactly a decade since the first establishment of SPMRL, a research initiative unifying multiple research efforts to address the peculiar challenges of Statistical Parsing for Morphologically-Rich Languages (MRLs). Here we reflect on parsing MRLs in that decade, highlight the solutions and lessons learned for the architectural, modeling and lexical challenges in the pre-neural era, and argue that similar challenges re-emerge in neural architectures for MRLs. We then aim to offer a climax, suggesting that incorporating symbolic ideas proposed in SPMRL terms into nowadays neural architectures has the potential to push NLP for MRLs to a new level. We sketch a strategies for designing Neural Models for MRLs (NMRL), and showcase preliminary support for these strategies via investigating the task of multi-tagging in Hebrew, a morphologically-rich, high-fusion, language.

Guiding Variational Response Generator to Exploit Persona

Bowen Wu, Mengyuan Li, Zongsheng Wang, Yifu Chen, Derek F. Wong, Qihang Feng, Junhong Huang and Baoxun Wang

Leveraging persona information of users in Neural Response Generators (NRG) to perform personalized conversations has been considered as an attractive and important topic in the research of conversational agents over the past few years. Despite of the promising progress achieved by recent studies in this field, persona information tends to be incorporated into neural networks in the form of user embeddings, with the expectation that the persona can be involved via End-to-End learning. This paper proposes to adopt the personality-related characteristics of human conversations into variational response generators, by designing a specific conditional variational autoencoder based deep model with two new regularization terms employed to the loss function, so as to guide the optimization towards the direction of generating both persona-aware and relevant responses. Besides, to reasonably evaluate the performances of various persona modeling approaches, this paper further presents three direct persona-oriented metrics from different perspectives. The experimental results have shown that our proposed methodology can notably improve the performance of persona-aware response generation, and the metrics are reasonable to evaluate the results.

Unsupervised Cross-lingual Representation Learning at Scale

Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov

This paper shows that pretraining multilingual language models at scale leads to significant performance gains for a wide range of cross-lingual transfer tasks. We train a Transformer-based masked language model on one hundred languages, using more than two terabytes of filtered CommonCrawl data. Our model, dubbed XLM-R, significantly outperforms multilingual BERT (mBERT) on a variety of cross-lingual benchmarks, including +14.6% average accuracy on XNLI, +13% average F1 score on MLQA, and +2.4% F1 score on NER. XLM-R performs particularly well on low-resource languages, improving 15.7% in XNLI accuracy for Swahili and 11.4% for Urdu over previous XLM models. We also present a detailed empirical analysis of the key factors that are required to achieve these gains, including the trade-offs between (1) positive transfer and capacity dilution and (2) the performance of high and low resource languages at scale. Finally, we show, for the first time, the possibility of multilingual modeling without sacrificing per-language performance; XLM-R is very competitive with strong monolingual models on the GLUE and XNLI benchmarks. We will make our code and models publicly available.

PLATO: Pre-trained Dialogue Generation Model with Discrete Latent Variable

Siqi Bao, Huang He, Fan Wang, Hua Wu and Haifeng Wang

Pre-training models have been proved effective for a wide range of natural language processing tasks. Inspired by this, we propose a novel dialogue generation pre-training framework to support various kinds of conversations, including chit-chat, knowledge grounded dialogues, and conversational question answering. In this framework, we adopt flexible attention mechanisms to fully leverage the bi-directional context and the uni-directional characteristic of language generation. We also introduce discrete latent variables to tackle the inherent one-to-many mapping problem in response generation. Two reciprocal tasks of response generation and latent act recognition are designed and carried out simultaneously within a shared network. Comprehensive experiments on three publicly available datasets verify the effectiveness and superiority of the proposed framework.

Hyperbolic Capsule Networks for Multi-Label Classification

Boli Chen, Xin Huang, Lin Xiao and Liping Jing

Although deep neural networks are effective at extracting high-level features, classification methods usually encode an input into a vector representation via simple feature aggregation operations (e.g. pooling). Such operations limit the performance. For instance, a multi-label document may contain several concepts. In this case, one vector can not sufficiently capture its salient and discriminative content. Thus, we propose Hyperbolic Capsule Networks (HyperCaps) for Multi-Label Classification (MLC), which have two merits. First, hyperbolic capsules are designed to capture fine-grained document information for each label, which has the ability to characterize complicated structures among labels and documents. Second, Hyperbolic Dynamic Routing (HDR) is introduced to aggregate hyperbolic capsules in a label-aware manner, so that the label-level discriminative information can be preserved along the depth of neural networks. To efficiently handle large-scale MLC datasets, we additionally present a new routing method to adaptively adjust the capsule number during routing. Extensive experiments are conducted on four benchmark datasets. Compared with the state-of-the-art methods, HyperCaps significantly improves the performance of MLC especially on tail labels.

Fine-Grained Analysis of Cross-Linguistic Syntactic Divergences

Dmitry Nikolaev, Ofir Arviv, Taelin Karidi, Neta Kenneth, Veronika Mitnik, Lilja Maria Saeboe and Omri Abend

The patterns in which the syntax of different languages converges and diverges are often used to inform work on cross-lingual transfer. Nevertheless, little empirical work has been done on quantifying the prevalence of different syntactic divergences across language pairs. We propose a framework for extracting divergence patterns for any language pair from a parallel corpus, building on Universal Dependencies. We show that our framework provides a detailed picture of cross-language divergences, generalizes previous approaches, and lends itself to full automation. We further present a novel dataset, a manually word-aligned subset of the Parallel UD corpus in five languages, and use it to perform a detailed corpus study. We demonstrate the usefulness of the resulting analysis by showing that it can help account for performance patterns of a cross-lingual parser.