paper_id
stringlengths
15
18
title
stringlengths
22
143
paper_url
stringlengths
41
44
authors
listlengths
1
16
abstract
large_stringlengths
439
1.79k
anthology_id
stringlengths
15
18
doi
stringlengths
27
30
award
stringclasses
3 values
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
2021.acl-short.30
Anchor-based Bilingual Word Embeddings for Low-Resource Languages
https://aclanthology.org/2021.acl-short.30/
[ "Tobias Eder", "Viktor Hangya", "Alexander Fraser" ]
Good quality monolingual word embeddings (MWEs) can be built for languages which have large amounts of unlabeled text. MWEs can be aligned to bilingual spaces using only a few thousand word translation pairs. For low resource languages training MWEs monolingually results in MWEs of poor quality, and thus poor bilingual...
2021.acl-short.30
10.18653/v1/2021.acl-short.30
null
2010.12627
title_snapshot
2021.acl-short.31
Multilingual Agreement for Multilingual Neural Machine Translation
https://aclanthology.org/2021.acl-short.31/
[ "Jian Yang", "Yuwei Yin", "Shuming Ma", "Haoyang Huang", "Dongdong Zhang", "Zhoujun Li", "Furu Wei" ]
Although multilingual neural machine translation (MNMT) enables multiple language translations, the training process is based on independent multilingual objectives. Most multilingual models can not explicitly exploit different language pairs to assist each other, ignoring the relationships among them. In this work, we...
2021.acl-short.31
10.18653/v1/2021.acl-short.31
null
null
null
2021.acl-short.32
Higher-order Derivatives of Weighted Finite-state Machines
https://aclanthology.org/2021.acl-short.32/
[ "Ran Zmigrod", "Tim Vieira", "Ryan Cotterell" ]
Weighted finite-state machines are a fundamental building block of NLP systems. They have withstood the test of time—from their early use in noisy channel models in the 1990s up to modern-day neurally parameterized conditional random fields. This work examines the computation of higher-order derivatives with respect to...
2021.acl-short.32
10.18653/v1/2021.acl-short.32
null
2106.00749
title_snapshot
2021.acl-short.33
Reinforcement Learning for Abstractive Question Summarization with Question-aware Semantic Rewards
https://aclanthology.org/2021.acl-short.33/
[ "Shweta Yadav", "Deepak Gupta", "Asma Ben Abacha", "Dina Demner-Fushman" ]
The growth of online consumer health questions has led to the necessity for reliable and accurate question answering systems. A recent study showed that manual summarization of consumer health questions brings significant improvement in retrieving relevant answers. However, the automatic summarization of long questions...
2021.acl-short.33
10.18653/v1/2021.acl-short.33
null
2107.00176
title_snapshot
2021.acl-short.34
A Semantics-aware Transformer Model of Relation Linking for Knowledge Base Question Answering
https://aclanthology.org/2021.acl-short.34/
[ "Tahira Naseem", "Srinivas Ravishankar", "Nandana Mihindukulasooriya", "Ibrahim Abdelaziz", "Young-Suk Lee", "Pavan Kapanipathi", "Salim Roukos", "Alfio Gliozzo", "Alexander Gray" ]
Relation linking is a crucial component of Knowledge Base Question Answering systems. Existing systems use a wide variety of heuristics, or ensembles of multiple systems, heavily relying on the surface question text. However, the explicit semantic parse of the question is a rich source of relation information that is n...
2021.acl-short.34
10.18653/v1/2021.acl-short.34
null
null
null
2021.acl-short.35
Neural Retrieval for Question Answering with Cross-Attention Supervised Data Augmentation
https://aclanthology.org/2021.acl-short.35/
[ "Yinfei Yang", "Ning Jin", "Kuo Lin", "Mandy Guo", "Daniel Cer" ]
Early fusion models with cross-attention have shown better-than-human performance on some question answer benchmarks, while it is a poor fit for retrieval since it prevents pre-computation of the answer representations. We present a supervised data mining method using an accurate early fusion model to improve the train...
2021.acl-short.35
10.18653/v1/2021.acl-short.35
null
2009.13815
title_snapshot
2021.acl-short.36
Enhancing Descriptive Image Captioning with Natural Language Inference
https://aclanthology.org/2021.acl-short.36/
[ "Zhan Shi", "Hui Liu", "Xiaodan Zhu" ]
Generating descriptive sentences that convey non-trivial, detailed, and salient information about images is an important goal of image captioning. In this paper we propose a novel approach to encourage captioning models to produce more detailed captions using natural language inference, based on the motivation that, am...
2021.acl-short.36
10.18653/v1/2021.acl-short.36
null
null
null
2021.acl-short.37
MOLEMAN: Mention-Only Linking of Entities with a Mention Annotation Network
https://aclanthology.org/2021.acl-short.37/
[ "Nicholas FitzGerald", "Dan Bikel", "Jan Botha", "Daniel Gillick", "Tom Kwiatkowski", "Andrew McCallum" ]
We present an instance-based nearest neighbor approach to entity linking. In contrast to most prior entity retrieval systems which represent each entity with a single vector, we build a contextualized mention-encoder that learns to place similar mentions of the same entity closer in vector space than mentions of differ...
2021.acl-short.37
10.18653/v1/2021.acl-short.37
null
2106.07352
title_snapshot
2021.acl-short.38
eMLM: A New Pre-training Objective for Emotion Related Tasks
https://aclanthology.org/2021.acl-short.38/
[ "Tiberiu Sosea", "Cornelia Caragea" ]
BERT has been shown to be extremely effective on a wide variety of natural language processing tasks, including sentiment analysis and emotion detection. However, the proposed pretraining objectives of BERT do not induce any sentiment or emotion-specific biases into the model. In this paper, we present Emotion Masked L...
2021.acl-short.38
10.18653/v1/2021.acl-short.38
null
null
null
2021.acl-short.39
On Positivity Bias in Negative Reviews
https://aclanthology.org/2021.acl-short.39/
[ "Madhusudhan Aithal", "Chenhao Tan" ]
Prior work has revealed that positive words occur more frequently than negative words in human expressions, which is typically attributed to positivity bias, a tendency for people to report positive views of reality. But what about the language used in negative reviews? Consistent with prior work, we show that English ...
2021.acl-short.39
10.18653/v1/2021.acl-short.39
null
2106.12056
title_snapshot
2021.acl-short.40
PRAL: A Tailored Pre-Training Model for Task-Oriented Dialog Generation
https://aclanthology.org/2021.acl-short.40/
[ "Jing Gu", "Qingyang Wu", "Chongruo Wu", "Weiyan Shi", "Zhou Yu" ]
Large pre-trained language generation models such as GPT-2 have demonstrated their effectiveness as language priors by reaching state-of-the-art results in various language generation tasks. However, the performance of pre-trained models on task-oriented dialog tasks is still under-explored. We propose a Pre-trainedRol...
2021.acl-short.40
10.18653/v1/2021.acl-short.40
null
2004.13835
title_judge
2021.acl-short.41
ROPE: Reading Order Equivariant Positional Encoding for Graph-based Document Information Extraction
https://aclanthology.org/2021.acl-short.41/
[ "Chen-Yu Lee", "Chun-Liang Li", "Chu Wang", "Renshen Wang", "Yasuhisa Fujii", "Siyang Qin", "Ashok Popat", "Tomas Pfister" ]
Natural reading orders of words are crucial for information extraction from form-like documents. Despite recent advances in Graph Convolutional Networks (GCNs) on modeling spatial layout patterns of documents, they have limited ability to capture reading orders of given word-level node representations in a graph. We pr...
2021.acl-short.41
10.18653/v1/2021.acl-short.41
null
2106.10786
title_snapshot
2021.acl-short.42
Zero-shot Event Extraction via Transfer Learning: Challenges and Insights
https://aclanthology.org/2021.acl-short.42/
[ "Qing Lyu", "Hongming Zhang", "Elior Sulem", "Dan Roth" ]
Event extraction has long been a challenging task, addressed mostly with supervised methods that require expensive annotation and are not extensible to new event ontologies. In this work, we explore the possibility of zero-shot event extraction by formulating it as a set of Textual Entailment (TE) and/or Question Answe...
2021.acl-short.42
10.18653/v1/2021.acl-short.42
null
null
null
2021.acl-short.43
Using Adversarial Attacks to Reveal the Statistical Bias in Machine Reading Comprehension Models
https://aclanthology.org/2021.acl-short.43/
[ "Jieyu Lin", "Jiajie Zou", "Nai Ding" ]
Pre-trained language models have achieved human-level performance on many Machine Reading Comprehension (MRC) tasks, but it remains unclear whether these models truly understand language or answer questions by exploiting statistical biases in datasets. Here, we demonstrate a simple yet effective method to attack MRC mo...
2021.acl-short.43
10.18653/v1/2021.acl-short.43
null
2105.11136
title_snapshot
2021.acl-short.44
Quantifying and Avoiding Unfair Qualification Labour in Crowdsourcing
https://aclanthology.org/2021.acl-short.44/
[ "Jonathan K. Kummerfeld" ]
Extensive work has argued in favour of paying crowd workers a wage that is at least equivalent to the U.S. federal minimum wage. Meanwhile, research on collecting high quality annotations suggests using a qualification that requires workers to have previously completed a certain number of tasks. If most requesters who ...
2021.acl-short.44
10.18653/v1/2021.acl-short.44
null
2105.12762
title_snapshot
2021.acl-short.45
Men Are Elected, Women Are Married: Events Gender Bias on Wikipedia
https://aclanthology.org/2021.acl-short.45/
[ "Jiao Sun", "Nanyun Peng" ]
Human activities can be seen as sequences of events, which are crucial to understanding societies. Disproportional event distribution for different demographic groups can manifest and amplify social stereotypes, and potentially jeopardize the ability of members in some groups to pursue certain goals. In this paper, we ...
2021.acl-short.45
10.18653/v1/2021.acl-short.45
null
2106.01601
title_snapshot
2021.acl-short.46
Modeling Task-Aware MIMO Cardinality for Efficient Multilingual Neural Machine Translation
https://aclanthology.org/2021.acl-short.46/
[ "Hongfei Xu", "Qiuhui Liu", "Josef van Genabith", "Deyi Xiong" ]
Neural machine translation has achieved great success in bilingual settings, as well as in multilingual settings. With the increase of the number of languages, multilingual systems tend to underperform their bilingual counterparts. Model capacity has been found crucial for massively multilingual NMT to support language...
2021.acl-short.46
10.18653/v1/2021.acl-short.46
null
null
null
2021.acl-short.47
Adaptive Nearest Neighbor Machine Translation
https://aclanthology.org/2021.acl-short.47/
[ "Xin Zheng", "Zhirui Zhang", "Junliang Guo", "Shujian Huang", "Boxing Chen", "Weihua Luo", "Jiajun Chen" ]
kNN-MT, recently proposed by Khandelwal et al. (2020a), successfully combines pre-trained neural machine translation (NMT) model with token-level k-nearest-neighbor (kNN) retrieval to improve the translation accuracy. However, the traditional kNN algorithm used in kNN-MT simply retrieves a same number of nearest neighb...
2021.acl-short.47
10.18653/v1/2021.acl-short.47
null
2105.13022
title_snapshot
2021.acl-short.48
On Orthogonality Constraints for Transformers
https://aclanthology.org/2021.acl-short.48/
[ "Aston Zhang", "Alvin Chan", "Yi Tay", "Jie Fu", "Shuohang Wang", "Shuai Zhang", "Huajie Shao", "Shuochao Yao", "Roy Ka-Wei Lee" ]
Orthogonality constraints encourage matrices to be orthogonal for numerical stability. These plug-and-play constraints, which can be conveniently incorporated into model training, have been studied for popular architectures in natural language processing, such as convolutional neural networks and recurrent neural netwo...
2021.acl-short.48
10.18653/v1/2021.acl-short.48
null
null
null
2021.acl-short.49
Measuring and Improving BERT’s Mathematical Abilities by Predicting the Order of Reasoning.
https://aclanthology.org/2021.acl-short.49/
[ "Piotr Piękos", "Mateusz Malinowski", "Henryk Michalewski" ]
Imagine you are in a supermarket. You have two bananas in your basket and want to buy four apples. How many fruits do you have in total? This seemingly straightforward question can be challenging for data-driven language models, even if trained at scale. However, we would expect such generic language models to possess ...
2021.acl-short.49
10.18653/v1/2021.acl-short.49
null
2106.03921
title_snapshot
2021.acl-short.50
Happy Dance, Slow Clap: Using Reaction GIFs to Predict Induced Affect on Twitter
https://aclanthology.org/2021.acl-short.50/
[ "Boaz Shmueli", "Soumya Ray", "Lun-Wei Ku" ]
Datasets with induced emotion labels are scarce but of utmost importance for many NLP tasks. We present a new, automated method for collecting texts along with their induced reaction labels. The method exploits the online use of reaction GIFs, which capture complex affective states. We show how to augment the data with...
2021.acl-short.50
10.18653/v1/2021.acl-short.50
null
2105.09967
title_snapshot
2021.acl-short.51
Exploring Listwise Evidence Reasoning with T5 for Fact Verification
https://aclanthology.org/2021.acl-short.51/
[ "Kelvin Jiang", "Ronak Pradeep", "Jimmy Lin" ]
This work explores a framework for fact verification that leverages pretrained sequence-to-sequence transformer models for sentence selection and label prediction, two key sub-tasks in fact verification. Most notably, improving on previous pointwise aggregation approaches for label prediction, we take advantage of T5 u...
2021.acl-short.51
10.18653/v1/2021.acl-short.51
null
null
null
2021.acl-short.52
DefSent: Sentence Embeddings using Definition Sentences
https://aclanthology.org/2021.acl-short.52/
[ "Hayato Tsukagoshi", "Ryohei Sasano", "Koichi Takeda" ]
Sentence embedding methods using natural language inference (NLI) datasets have been successfully applied to various tasks. However, these methods are only available for limited languages due to relying heavily on the large NLI datasets. In this paper, we propose DefSent, a sentence embedding method that uses definitio...
2021.acl-short.52
10.18653/v1/2021.acl-short.52
null
2105.04339
title_snapshot
2021.acl-short.53
Discrete Cosine Transform as Universal Sentence Encoder
https://aclanthology.org/2021.acl-short.53/
[ "Nada Almarwani", "Mona Diab" ]
Modern sentence encoders are used to generate dense vector representations that capture the underlying linguistic characteristics for a sequence of words, including phrases, sentences, or paragraphs. These kinds of representations are ideal for training a classifier for an end task such as sentiment analysis, question ...
2021.acl-short.53
10.18653/v1/2021.acl-short.53
null
2106.00934
title_snapshot
2021.acl-short.54
AligNarr: Aligning Narratives on Movies
https://aclanthology.org/2021.acl-short.54/
[ "Paramita Mirza", "Mostafa Abouhamra", "Gerhard Weikum" ]
High-quality alignment between movie scripts and plot summaries is an asset for learning to summarize stories and to generate dialogues. The alignment task is challenging as scripts and summaries substantially differ in details and abstraction levels as well as in linguistic register. This paper addresses the alignment...
2021.acl-short.54
10.18653/v1/2021.acl-short.54
null
null
null
2021.acl-short.55
An Exploratory Analysis of Multilingual Word-Level Quality Estimation with Cross-Lingual Transformers
https://aclanthology.org/2021.acl-short.55/
[ "Tharindu Ranasinghe", "Constantin Orasan", "Ruslan Mitkov" ]
Most studies on word-level Quality Estimation (QE) of machine translation focus on language-specific models. The obvious disadvantages of these approaches are the need for labelled data for each language pair and the high cost required to maintain several language-specific models. To overcome these problems, we explore...
2021.acl-short.55
10.18653/v1/2021.acl-short.55
null
2106.00143
title_snapshot
2021.acl-short.56
Exploration and Exploitation: Two Ways to Improve Chinese Spelling Correction Models
https://aclanthology.org/2021.acl-short.56/
[ "Chong Li", "Cenyuan Zhang", "Xiaoqing Zheng", "Xuanjing Huang" ]
A sequence-to-sequence learning with neural networks has empirically proven to be an effective framework for Chinese Spelling Correction (CSC), which takes a sentence with some spelling errors as input and outputs the corrected one. However, CSC models may fail to correct spelling errors covered by the confusion sets, ...
2021.acl-short.56
10.18653/v1/2021.acl-short.56
null
2105.14813
title_snapshot
2021.acl-short.57
Training Adaptive Computation for Open-Domain Question Answering with Computational Constraints
https://aclanthology.org/2021.acl-short.57/
[ "Yuxiang Wu", "Pasquale Minervini", "Pontus Stenetorp", "Sebastian Riedel" ]
Adaptive Computation (AC) has been shown to be effective in improving the efficiency of Open-Domain Question Answering (ODQA) systems. However, the current AC approaches require tuning of all model parameters, and training state-of-the-art ODQA models requires significant computational resources that may not be availab...
2021.acl-short.57
10.18653/v1/2021.acl-short.57
null
2107.02102
title_snapshot
2021.acl-short.58
An Empirical Study on Adversarial Attack on NMT: Languages and Positions Matter
https://aclanthology.org/2021.acl-short.58/
[ "Zhiyuan Zeng", "Deyi Xiong" ]
In this paper, we empirically investigate adversarial attack on NMT from two aspects: languages (the source vs. the target language) and positions (front vs. rear). For autoregressive NMT models that generate target words from left to right, we observe that adversarial attack on the source language is more effective th...
2021.acl-short.58
10.18653/v1/2021.acl-short.58
null
null
null
2021.acl-short.59
OntoGUM: Evaluating Contextualized SOTA Coreference Resolution on 12 More Genres
https://aclanthology.org/2021.acl-short.59/
[ "Yilun Zhu", "Sameer Pradhan", "Amir Zeldes" ]
SOTA coreference resolution produces increasingly impressive scores on the OntoNotes benchmark. However lack of comparable data following the same scheme for more genres makes it difficult to evaluate generalizability to open domain data. This paper provides a dataset and comprehensive evaluation showing that the lates...
2021.acl-short.59
10.18653/v1/2021.acl-short.59
null
2106.00933
title_snapshot
2021.acl-short.60
In Factuality: Efficient Integration of Relevant Facts for Visual Question Answering
https://aclanthology.org/2021.acl-short.60/
[ "Peter Vickers", "Nikolaos Aletras", "Emilio Monti", "Loïc Barrault" ]
Visual Question Answering (VQA) methods aim at leveraging visual input to answer questions that may require complex reasoning over entities. Current models are trained on labelled data that may be insufficient to learn complex knowledge representations. In this paper, we propose a new method to enhance the reasoning ca...
2021.acl-short.60
10.18653/v1/2021.acl-short.60
null
null
null
2021.acl-short.61
Zero-shot Fact Verification by Claim Generation
https://aclanthology.org/2021.acl-short.61/
[ "Liangming Pan", "Wenhu Chen", "Wenhan Xiong", "Min-Yen Kan", "William Yang Wang" ]
Neural models for automated fact verification have achieved promising results thanks to the availability of large, human-annotated datasets. However, for each new domain that requires fact verification, creating a dataset by manually writing claims and linking them to their supporting evidence is expensive. We develop ...
2021.acl-short.61
10.18653/v1/2021.acl-short.61
null
2105.14682
title_snapshot
2021.acl-short.62
Thank you BART! Rewarding Pre-Trained Models Improves Formality Style Transfer
https://aclanthology.org/2021.acl-short.62/
[ "Huiyuan Lai", "Antonio Toral", "Malvina Nissim" ]
Scarcity of parallel data causes formality style transfer models to have scarce success in preserving content. We show that fine-tuning pre-trained language (GPT-2) and sequence-to-sequence (BART) models boosts content preservation, and that this is possible even with limited amounts of parallel data. Augmenting these ...
2021.acl-short.62
10.18653/v1/2021.acl-short.62
null
2105.06947
title_snapshot
2021.acl-short.63
Deep Context- and Relation-Aware Learning for Aspect-based Sentiment Analysis
https://aclanthology.org/2021.acl-short.63/
[ "Shinhyeok Oh", "Dongyub Lee", "Taesun Whang", "IlNam Park", "Seo Gaeun", "EungGyun Kim", "Harksoo Kim" ]
Existing works for aspect-based sentiment analysis (ABSA) have adopted a unified approach, which allows the interactive relations among subtasks. However, we observe that these methods tend to predict polarities based on the literal meaning of aspect and opinion terms and mainly consider relations implicitly among subt...
2021.acl-short.63
10.18653/v1/2021.acl-short.63
null
2106.03806
title_snapshot
2021.acl-short.64
Towards Generative Aspect-Based Sentiment Analysis
https://aclanthology.org/2021.acl-short.64/
[ "Wenxuan Zhang", "Xin Li", "Yang Deng", "Lidong Bing", "Wai Lam" ]
Aspect-based sentiment analysis (ABSA) has received increasing attention recently. Most existing work tackles ABSA in a discriminative manner, designing various task-specific classification networks for the prediction. Despite their effectiveness, these methods ignore the rich label semantics in ABSA problems and requi...
2021.acl-short.64
10.18653/v1/2021.acl-short.64
null
null
null
2021.acl-short.65
Bilingual Mutual Information Based Adaptive Training for Neural Machine Translation
https://aclanthology.org/2021.acl-short.65/
[ "Yangyifan Xu", "Yijin Liu", "Fandong Meng", "Jiajun Zhang", "Jinan Xu", "Jie Zhou" ]
Recently, token-level adaptive training has achieved promising improvement in machine translation, where the cross-entropy loss function is adjusted by assigning different training weights to different tokens, in order to alleviate the token imbalance problem. However, previous approaches only use static word frequency...
2021.acl-short.65
10.18653/v1/2021.acl-short.65
null
2105.12523
title_snapshot
2021.acl-short.66
Continual Learning for Task-oriented Dialogue System with Iterative Network Pruning, Expanding and Masking
https://aclanthology.org/2021.acl-short.66/
[ "Binzong Geng", "Fajie Yuan", "Qiancheng Xu", "Ying Shen", "Ruifeng Xu", "Min Yang" ]
This ability to learn consecutive tasks without forgetting how to perform previously trained problems is essential for developing an online dialogue system. This paper proposes an effective continual learning method for the task-oriented dialogue system with iterative network pruning, expanding, and masking (TPEM), whi...
2021.acl-short.66
10.18653/v1/2021.acl-short.66
null
2107.08173
title_snapshot
2021.acl-short.67
TIMERS: Document-level Temporal Relation Extraction
https://aclanthology.org/2021.acl-short.67/
[ "Puneet Mathur", "Rajiv Jain", "Franck Dernoncourt", "Vlad Morariu", "Quan Hung Tran", "Dinesh Manocha" ]
We present TIMERS - a TIME, Rhetorical and Syntactic-aware model for document-level temporal relation classification in the English language. Our proposed method leverages rhetorical discourse features and temporal arguments from semantic role labels, in addition to traditional local syntactic features, trained through...
2021.acl-short.67
10.18653/v1/2021.acl-short.67
null
null
null
2021.acl-short.68
Improving Arabic Diacritization with Regularized Decoding and Adversarial Training
https://aclanthology.org/2021.acl-short.68/
[ "Han Qin", "Guimin Chen", "Yuanhe Tian", "Yan Song" ]
Arabic diacritization is a fundamental task for Arabic language processing. Previous studies have demonstrated that automatically generated knowledge can be helpful to this task. However, these studies regard the auto-generated knowledge instances as gold references, which limits their effectiveness since such knowledg...
2021.acl-short.68
10.18653/v1/2021.acl-short.68
null
null
null
2021.acl-short.69
When is Char Better Than Subword: A Systematic Study of Segmentation Algorithms for Neural Machine Translation
https://aclanthology.org/2021.acl-short.69/
[ "Jiahuan Li", "Yutong Shen", "Shujian Huang", "Xinyu Dai", "Jiajun Chen" ]
Subword segmentation algorithms have been a de facto choice when building neural machine translation systems. However, most of them need to learn a segmentation model based on some heuristics, which may produce sub-optimal segmentation. This can be problematic in some scenarios when the target language has rich morphol...
2021.acl-short.69
10.18653/v1/2021.acl-short.69
null
null
null
2021.acl-short.70
More than Text: Multi-modal Chinese Word Segmentation
https://aclanthology.org/2021.acl-short.70/
[ "Dong Zhang", "Zheng Hu", "Shoushan Li", "Hanqian Wu", "Qiaoming Zhu", "Guodong Zhou" ]
Chinese word segmentation (CWS) is undoubtedly an important basic task in natural language processing. Previous works only focus on the textual modality, but there are often audio and video utterances (such as news broadcast and face-to-face dialogues), where textual, acoustic and visual modalities normally exist. To t...
2021.acl-short.70
10.18653/v1/2021.acl-short.70
null
null
null
2021.acl-short.71
A Mixture-of-Experts Model for Antonym-Synonym Discrimination
https://aclanthology.org/2021.acl-short.71/
[ "Zhipeng Xie", "Nan Zeng" ]
Discrimination between antonyms and synonyms is an important and challenging NLP task. Antonyms and synonyms often share the same or similar contexts and thus are hard to make a distinction. This paper proposes two underlying hypotheses and employs the mixture-of-experts framework as a solution. It works on the basis o...
2021.acl-short.71
10.18653/v1/2021.acl-short.71
null
null
null
2021.acl-short.72
Learning Domain-Specialised Representations for Cross-Lingual Biomedical Entity Linking
https://aclanthology.org/2021.acl-short.72/
[ "Fangyu Liu", "Ivan Vulić", "Anna Korhonen", "Nigel Collier" ]
Injecting external domain-specific knowledge (e.g., UMLS) into pretrained language models (LMs) advances their capability to handle specialised in-domain tasks such as biomedical entity linking (BEL). However, such abundant expert knowledge is available only for a handful of languages (e.g., English). In this work, by ...
2021.acl-short.72
10.18653/v1/2021.acl-short.72
null
2105.14398
title_snapshot
2021.acl-short.73
A Cluster-based Approach for Improving Isotropy in Contextual Embedding Space
https://aclanthology.org/2021.acl-short.73/
[ "Sara Rajaee", "Mohammad Taher Pilehvar" ]
The representation degeneration problem in Contextual Word Representations (CWRs) hurts the expressiveness of the embedding space by forming an anisotropic cone where even unrelated words have excessively positive correlations. Existing techniques for tackling this issue require a learning process to re-train models wi...
2021.acl-short.73
10.18653/v1/2021.acl-short.73
null
2106.01183
title_snapshot
2021.acl-short.74
Unsupervised Enrichment of Persona-grounded Dialog with Background Stories
https://aclanthology.org/2021.acl-short.74/
[ "Bodhisattwa Prasad Majumder", "Taylor Berg-Kirkpatrick", "Julian McAuley", "Harsh Jhamtani" ]
Humans often refer to personal narratives, life experiences, and events to make a conversation more engaging and rich. While persona-grounded dialog models are able to generate responses that follow a given persona, they often miss out on stating detailed experiences or events related to a persona, often leaving conver...
2021.acl-short.74
10.18653/v1/2021.acl-short.74
null
2106.08364
title_snapshot
2021.acl-short.75
Beyond Laurel/Yanny: An Autoencoder-Enabled Search for Polyperceivable Audio
https://aclanthology.org/2021.acl-short.75/
[ "Kartik Chandra", "Chuma Kabaghe", "Gregory Valiant" ]
The famous “laurel/yanny” phenomenon references an audio clip that elicits dramatically different responses from different listeners. For the original clip, roughly half the population hears the word “laurel,” while the other half hears “yanny.” How common are such “polyperceivable” audio clips? In this paper we apply ...
2021.acl-short.75
10.18653/v1/2021.acl-short.75
null
null
null
2021.acl-short.76
Don’t Let Discourse Confine Your Model: Sequence Perturbations for Improved Event Language Models
https://aclanthology.org/2021.acl-short.76/
[ "Mahnaz Koupaee", "Greg Durrett", "Nathanael Chambers", "Niranjan Balasubramanian" ]
Event language models represent plausible sequences of events. Most existing approaches train autoregressive models on text, which successfully capture event co-occurrence but unfortunately constrain the model to follow the discourse order in which events are presented. Other domains may employ different discourse orde...
2021.acl-short.76
10.18653/v1/2021.acl-short.76
null
null
null
2021.acl-short.77
The Curse of Dense Low-Dimensional Information Retrieval for Large Index Sizes
https://aclanthology.org/2021.acl-short.77/
[ "Nils Reimers", "Iryna Gurevych" ]
Information Retrieval using dense low-dimensional representations recently became popular and showed out-performance to traditional sparse-representations like BM25. However, no previous work investigated how dense representations perform with large index sizes. We show theoretically and empirically that the performanc...
2021.acl-short.77
10.18653/v1/2021.acl-short.77
null
2012.14210
title_snapshot
2021.acl-short.78
Cross-lingual Text Classification with Heterogeneous Graph Neural Network
https://aclanthology.org/2021.acl-short.78/
[ "Ziyun Wang", "Xuan Liu", "Peiji Yang", "Shixing Liu", "Zhisheng Wang" ]
Cross-lingual text classification aims at training a classifier on the source language and transferring the knowledge to target languages, which is very useful for low-resource languages. Recent multilingual pretrained language models (mPLM) achieve impressive results in cross-lingual classification tasks, but rarely c...
2021.acl-short.78
10.18653/v1/2021.acl-short.78
null
2105.11246
title_snapshot
2021.acl-short.79
Towards more equitable question answering systems: How much more data do you need?
https://aclanthology.org/2021.acl-short.79/
[ "Arnab Debnath", "Navid Rajabi", "Fardina Fathmiul Alam", "Antonios Anastasopoulos" ]
Question answering (QA) in English has been widely explored, but multilingual datasets are relatively new, with several methods attempting to bridge the gap between high- and low-resourced languages using data augmentation through translation and cross-lingual transfer. In this project we take a step back and study whi...
2021.acl-short.79
10.18653/v1/2021.acl-short.79
null
2105.14115
title_snapshot
2021.acl-short.80
Embedding Time Differences in Context-sensitive Neural Networks for Learning Time to Event
https://aclanthology.org/2021.acl-short.80/
[ "Nazanin Dehghani", "Hassan Hajipoor", "Hadi Amiri" ]
We propose an effective context-sensitive neural model for time to event (TTE) prediction task, which aims to predict the amount of time to/from the occurrence of given events in streaming content. We investigate this problem in the context of a multi-task learning framework, which we enrich with time difference embedd...
2021.acl-short.80
10.18653/v1/2021.acl-short.80
null
null
null
2021.acl-short.81
Improving Compositional Generalization in Classification Tasks via Structure Annotations
https://aclanthology.org/2021.acl-short.81/
[ "Juyong Kim", "Pradeep Ravikumar", "Joshua Ainslie", "Santiago Ontañón" ]
Compositional generalization is the ability to generalize systematically to a new data distribution by combining known components. Although humans seem to have a great ability to generalize compositionally, state-of-the-art neural models struggle to do so. In this work, we study compositional generalization in classifi...
2021.acl-short.81
10.18653/v1/2021.acl-short.81
null
2106.10434
title_snapshot
2021.acl-short.82
Learning to Generate Task-Specific Adapters from Task Description
https://aclanthology.org/2021.acl-short.82/
[ "Qinyuan Ye", "Xiang Ren" ]
Pre-trained text-to-text transformers such as BART have achieved impressive performance across a range of NLP tasks. Recent study further shows that they can learn to generalize to novel tasks, by including task descriptions as part of the source sequence and training the model with (source, target) examples. At test t...
2021.acl-short.82
10.18653/v1/2021.acl-short.82
null
2101.00420
title_snapshot
2021.acl-short.83
QA-Driven Zero-shot Slot Filling with Weak Supervision Pretraining
https://aclanthology.org/2021.acl-short.83/
[ "Xinya Du", "Luheng He", "Qi Li", "Dian Yu", "Panupong Pasupat", "Yuan Zhang" ]
Slot-filling is an essential component for building task-oriented dialog systems. In this work, we focus on the zero-shot slot-filling problem, where the model needs to predict slots and their values, given utterances from new domains without training on the target domain. Prior methods directly encode slot description...
2021.acl-short.83
10.18653/v1/2021.acl-short.83
null
null
null
2021.acl-short.84
Domain-Adaptive Pretraining Methods for Dialogue Understanding
https://aclanthology.org/2021.acl-short.84/
[ "Han Wu", "Kun Xu", "Linfeng Song", "Lifeng Jin", "Haisong Zhang", "Linqi Song" ]
Language models like BERT and SpanBERT pretrained on open-domain data have obtained impressive gains on various NLP tasks. In this paper, we probe the effectiveness of domain-adaptive pretraining objectives on downstream tasks. In particular, three objectives, including a novel objective focusing on modeling predicate-...
2021.acl-short.84
10.18653/v1/2021.acl-short.84
null
2105.13665
title_snapshot
2021.acl-short.85
Targeting the Benchmark: On Methodology in Current Natural Language Processing Research
https://aclanthology.org/2021.acl-short.85/
[ "David Schlangen" ]
It has become a common pattern in our field: One group introduces a language task, exemplified by a dataset, which they argue is challenging enough to serve as a benchmark. They also provide a baseline model for it, which then soon is improved upon by other groups. Often, research efforts then move on, and the pattern ...
2021.acl-short.85
10.18653/v1/2021.acl-short.85
null
2007.04792
title_snapshot
2021.acl-short.86
X-Fact: A New Benchmark Dataset for Multilingual Fact Checking
https://aclanthology.org/2021.acl-short.86/
[ "Ashim Gupta", "Vivek Srikumar" ]
In this work, we introduce : the largest publicly available multilingual dataset for factual verification of naturally existing real-world claims. The dataset contains short statements in 25 languages and is labeled for veracity by expert fact-checkers. The dataset includes a multilingual evaluation benchmark that meas...
2021.acl-short.86
10.18653/v1/2021.acl-short.86
null
2106.09248
title_snapshot
2021.acl-short.87
nmT5 - Is parallel data still relevant for pre-training massively multilingual language models?
https://aclanthology.org/2021.acl-short.87/
[ "Mihir Kale", "Aditya Siddhant", "Rami Al-Rfou", "Linting Xue", "Noah Constant", "Melvin Johnson" ]
Recently, mT5 - a massively multilingual version of T5 - leveraged a unified text-to-text format to attain state-of-the-art results on a wide variety of multilingual NLP tasks. In this paper, we investigate the impact of incorporating parallel data into mT5 pre-training. We find that multi-tasking language modeling wit...
2021.acl-short.87
10.18653/v1/2021.acl-short.87
null
2106.02171
title_snapshot
2021.acl-short.88
Question Generation for Adaptive Education
https://aclanthology.org/2021.acl-short.88/
[ "Megha Srivastava", "Noah Goodman" ]
Intelligent and adaptive online education systems aim to make high-quality education available for a diverse range of students. However, existing systems usually depend on a pool of hand-made questions, limiting how fine-grained and open-ended they can be in adapting to individual students. We explore targeted question...
2021.acl-short.88
10.18653/v1/2021.acl-short.88
null
2106.04262
title_snapshot
2021.acl-short.89
A Simple Recipe for Multilingual Grammatical Error Correction
https://aclanthology.org/2021.acl-short.89/
[ "Sascha Rothe", "Jonathan Mallinson", "Eric Malmi", "Sebastian Krause", "Aliaksei Severyn" ]
This paper presents a simple recipe to trainstate-of-the-art multilingual Grammatical Error Correction (GEC) models. We achieve this by first proposing a language-agnostic method to generate a large number of synthetic examples. The second ingredient is to use large-scale multilingual language models (up to 11B paramet...
2021.acl-short.89
10.18653/v1/2021.acl-short.89
null
2106.03830
title_snapshot
2021.acl-short.90
Towards Visual Question Answering on Pathology Images
https://aclanthology.org/2021.acl-short.90/
[ "Xuehai He", "Zhuo Cai", "Wenlan Wei", "Yichen Zhang", "Luntian Mou", "Eric Xing", "Pengtao Xie" ]
Pathology imaging is broadly used for identifying the causes and effects of diseases or injuries. Given a pathology image, being able to answer questions about the clinical findings contained in the image is very important for medical decision making. In this paper, we aim to develop a pathological visual question answ...
2021.acl-short.90
10.18653/v1/2021.acl-short.90
null
null
null
2021.acl-short.91
Efficient Text-based Reinforcement Learning by Jointly Leveraging State and Commonsense Graph Representations
https://aclanthology.org/2021.acl-short.91/
[ "Keerthiram Murugesan", "Mattia Atzeni", "Pavan Kapanipathi", "Kartik Talamadupula", "Mrinmaya Sachan", "Murray Campbell" ]
Text-based games (TBGs) have emerged as useful benchmarks for evaluating progress at the intersection of grounded language understanding and reinforcement learning (RL). Recent work has proposed the use of external knowledge to improve the efficiency of RL agents for TBGs. In this paper, we posit that to act efficientl...
2021.acl-short.91
10.18653/v1/2021.acl-short.91
null
null
null
2021.acl-short.92
mTVR: Multilingual Moment Retrieval in Videos
https://aclanthology.org/2021.acl-short.92/
[ "Jie Lei", "Tamara Berg", "Mohit Bansal" ]
We introduce mTVR, a large-scale multilingual video moment retrieval dataset, containing 218K English and Chinese queries from 21.8K TV show video clips. The dataset is collected by extending the popular TVR dataset (in English) with paired Chinese queries and subtitles. Compared to existing moment retrieval datasets, ...
2021.acl-short.92
10.18653/v1/2021.acl-short.92
null
2108.00061
title_snapshot
2021.acl-short.93
Explicitly Capturing Relations between Entity Mentions via Graph Neural Networks for Domain-specific Named Entity Recognition
https://aclanthology.org/2021.acl-short.93/
[ "Pei Chen", "Haibo Ding", "Jun Araki", "Ruihong Huang" ]
Named entity recognition (NER) is well studied for the general domain, and recent systems have achieved human-level performance for identifying common entity types. However, the NER performance is still moderate for specialized domains that tend to feature complicated contexts and jargonistic entity types. To address t...
2021.acl-short.93
10.18653/v1/2021.acl-short.93
null
null
null
2021.acl-short.94
Improving Lexically Constrained Neural Machine Translation with Source-Conditioned Masked Span Prediction
https://aclanthology.org/2021.acl-short.94/
[ "Gyubok Lee", "Seongjun Yang", "Edward Choi" ]
Accurate terminology translation is crucial for ensuring the practicality and reliability of neural machine translation (NMT) systems. To address this, lexically constrained NMT explores various methods to ensure pre-specified words and phrases appear in the translation output. However, in many cases, those methods are...
2021.acl-short.94
10.18653/v1/2021.acl-short.94
null
2105.05498
title_snapshot
2021.acl-short.95
Quotation Recommendation and Interpretation Based on Transformation from Queries to Quotations
https://aclanthology.org/2021.acl-short.95/
[ "Lingzhi Wang", "Xingshan Zeng", "Kam-Fai Wong" ]
To help individuals express themselves better, quotation recommendation is receiving growing attention. Nevertheless, most prior efforts focus on modeling quotations and queries separately and ignore the relationship between the quotations and the queries. In this work, we introduce a transformation matrix that directl...
2021.acl-short.95
10.18653/v1/2021.acl-short.95
null
2105.14189
title_snapshot
2021.acl-short.96
Pre-training is a Hot Topic: Contextualized Document Embeddings Improve Topic Coherence
https://aclanthology.org/2021.acl-short.96/
[ "Federico Bianchi", "Silvia Terragni", "Dirk Hovy" ]
Topic models extract groups of words from documents, whose interpretation as a topic hopefully allows for a better understanding of the data. However, the resulting word groups are often not coherent, making them harder to interpret. Recently, neural topic models have shown improvements in overall coherence. Concurrent...
2021.acl-short.96
10.18653/v1/2021.acl-short.96
null
2004.03974
title_snapshot
2021.acl-short.97
Input Representations for Parsing Discourse Representation Structures: Comparing English with Chinese
https://aclanthology.org/2021.acl-short.97/
[ "Chunliu Wang", "Rik van Noord", "Arianna Bisazza", "Johan Bos" ]
Neural semantic parsers have obtained acceptable results in the context of parsing DRSs (Discourse Representation Structures). In particular models with character sequences as input showed remarkable performance for English. But how does this approach perform on languages with a different writing system, like Chinese, ...
2021.acl-short.97
10.18653/v1/2021.acl-short.97
null
null
null
2021.acl-short.98
Code Generation from Natural Language with Less Prior Knowledge and More Monolingual Data
https://aclanthology.org/2021.acl-short.98/
[ "Sajad Norouzi", "Keyi Tang", "Yanshuai Cao" ]
Training datasets for semantic parsing are typically small due to the higher expertise required for annotation than most other NLP tasks. As a result, models for this application usually need additional prior knowledge to be built into the architecture or algorithm. The increased dependency on human experts hinders aut...
2021.acl-short.98
10.18653/v1/2021.acl-short.98
null
2101.00259
title_judge
2021.acl-short.99
Issues with Entailment-based Zero-shot Text Classification
https://aclanthology.org/2021.acl-short.99/
[ "Tingting Ma", "Jin-Ge Yao", "Chin-Yew Lin", "Tiejun Zhao" ]
The general format of natural language inference (NLI) makes it tempting to be used for zero-shot text classification by casting any target label into a sentence of hypothesis and verifying whether or not it could be entailed by the input, aiming at generic classification applicable on any specified label space. In thi...
2021.acl-short.99
10.18653/v1/2021.acl-short.99
null
null
null
2021.acl-short.100
Neural-Symbolic Commonsense Reasoner with Relation Predictors
https://aclanthology.org/2021.acl-short.100/
[ "Farhad Moghimifar", "Lizhen Qu", "Terry Yue Zhuo", "Gholamreza Haffari", "Mahsa Baktashmotlagh" ]
Commonsense reasoning aims to incorporate sets of commonsense facts, retrieved from Commonsense Knowledge Graphs (CKG), to draw conclusion about ordinary situations. The dynamic nature of commonsense knowledge postulates models capable of performing multi-hop reasoning over new situations. This feature also results in ...
2021.acl-short.100
10.18653/v1/2021.acl-short.100
null
2105.06717
title_snapshot
2021.acl-short.101
What Motivates You? Benchmarking Automatic Detection of Basic Needs from Short Posts
https://aclanthology.org/2021.acl-short.101/
[ "Sanja Stajner", "Seren Yenikent", "Bilal Ghanem", "Marc Franco-Salvador" ]
According to the self-determination theory, the levels of satisfaction of three basic needs (competence, autonomy and relatedness) have implications on people’s everyday life and career. We benchmark the novel task of automatically detecting those needs on short posts in English, by modelling it as a ternary classifica...
2021.acl-short.101
10.18653/v1/2021.acl-short.101
null
null
null
2021.acl-short.102
Semantic Frame Induction using Masked Word Embeddings and Two-Step Clustering
https://aclanthology.org/2021.acl-short.102/
[ "Kosuke Yamada", "Ryohei Sasano", "Koichi Takeda" ]
Recent studies on semantic frame induction show that relatively high performance has been achieved by using clustering-based methods with contextualized word embeddings. However, there are two potential drawbacks to these methods: one is that they focus too much on the superficial information of the frame-evoking verb ...
2021.acl-short.102
10.18653/v1/2021.acl-short.102
null
2105.13466
title_snapshot
2021.acl-short.103
Lightweight Adapter Tuning for Multilingual Speech Translation
https://aclanthology.org/2021.acl-short.103/
[ "Hang Le", "Juan Pino", "Changhan Wang", "Jiatao Gu", "Didier Schwab", "Laurent Besacier" ]
Adapter modules were recently introduced as an efficient alternative to fine-tuning in NLP. Adapter tuning consists in freezing pre-trained parameters of a model and injecting lightweight modules between layers, resulting in the addition of only a small number of task-specific trainable parameters. While adapter tuning...
2021.acl-short.103
10.18653/v1/2021.acl-short.103
null
2106.01463
title_snapshot
2021.acl-short.104
Parameter Selection: Why We Should Pay More Attention to It
https://aclanthology.org/2021.acl-short.104/
[ "Jie-Jyun Liu", "Tsung-Han Yang", "Si-An Chen", "Chih-Jen Lin" ]
The importance of parameter selection in supervised learning is well known. However, due to the many parameter combinations, an incomplete or an insufficient procedure is often applied. This situation may cause misleading or confusing conclusions. In this opinion paper, through an intriguing example we point out that t...
2021.acl-short.104
10.18653/v1/2021.acl-short.104
null
2107.05393
title_snapshot
2021.acl-short.105
Distinct Label Representations for Few-Shot Text Classification
https://aclanthology.org/2021.acl-short.105/
[ "Sora Ohashi", "Junya Takayama", "Tomoyuki Kajiwara", "Yuki Arase" ]
Few-shot text classification aims to classify inputs whose label has only a few examples. Previous studies overlooked the semantic relevance between label representations. Therefore, they are easily confused by labels that are relevant. To address this problem, we propose a method that generates distinct label represen...
2021.acl-short.105
10.18653/v1/2021.acl-short.105
null
null
null
2021.acl-short.106
Learning to Solve NLP Tasks in an Incremental Number of Languages
https://aclanthology.org/2021.acl-short.106/
[ "Giuseppe Castellucci", "Simone Filice", "Danilo Croce", "Roberto Basili" ]
In real scenarios, a multilingual model trained to solve NLP tasks on a set of languages can be required to support new languages over time. Unfortunately, the straightforward retraining on a dataset containing annotated examples for all the languages is both expensive and time-consuming, especially when the number of ...
2021.acl-short.106
10.18653/v1/2021.acl-short.106
null
null
null
2021.acl-short.107
Hi-Transformer: Hierarchical Interactive Transformer for Efficient and Effective Long Document Modeling
https://aclanthology.org/2021.acl-short.107/
[ "Chuhan Wu", "Fangzhao Wu", "Tao Qi", "Yongfeng Huang" ]
Transformer is important for text modeling. However, it has difficulty in handling long documents due to the quadratic complexity with input text length. In order to handle this problem, we propose a hierarchical interactive Transformer (Hi-Transformer) for efficient and effective long document modeling. Hi-Transformer...
2021.acl-short.107
10.18653/v1/2021.acl-short.107
null
2106.01040
title_snapshot
2021.acl-short.108
Robust Transfer Learning with Pretrained Language Models through Adapters
https://aclanthology.org/2021.acl-short.108/
[ "Wenjuan Han", "Bo Pang", "Ying Nian Wu" ]
Transfer learning with large pretrained transformer-based language models like BERT has become a dominating approach for most NLP tasks. Simply fine-tuning those large language models on downstream tasks or combining it with task-specific pretraining is often not robust. In particular, the performance considerably vari...
2021.acl-short.108
10.18653/v1/2021.acl-short.108
null
2108.02340
title_snapshot
2021.acl-short.109
Embracing Ambiguity: Shifting the Training Target of NLI Models
https://aclanthology.org/2021.acl-short.109/
[ "Johannes Mario Meissner", "Napat Thumwanit", "Saku Sugawara", "Akiko Aizawa" ]
Natural Language Inference (NLI) datasets contain examples with highly ambiguous labels. While many research works do not pay much attention to this fact, several recent efforts have been made to acknowledge and embrace the existence of ambiguity, such as UNLI and ChaosNLI. In this paper, we explore the option of train...
2021.acl-short.109
10.18653/v1/2021.acl-short.109
null
2106.03020
title_snapshot
2021.acl-short.110
Modeling Discriminative Representations for Out-of-Domain Detection with Supervised Contrastive Learning
https://aclanthology.org/2021.acl-short.110/
[ "Zhiyuan Zeng", "Keqing He", "Yuanmeng Yan", "Zijun Liu", "Yanan Wu", "Hong Xu", "Huixing Jiang", "Weiran Xu" ]
Detecting Out-of-Domain (OOD) or unknown intents from user queries is essential in a task-oriented dialog system. A key challenge of OOD detection is to learn discriminative semantic features. Traditional cross-entropy loss only focuses on whether a sample is correctly classified, and does not explicitly distinguish th...
2021.acl-short.110
10.18653/v1/2021.acl-short.110
null
2105.14289
title_snapshot
2021.acl-short.111
Preview, Attend and Review: Schema-Aware Curriculum Learning for Multi-Domain Dialogue State Tracking
https://aclanthology.org/2021.acl-short.111/
[ "Yinpei Dai", "Hangyu Li", "Yongbin Li", "Jian Sun", "Fei Huang", "Luo Si", "Xiaodan Zhu" ]
Existing dialog state tracking (DST) models are trained with dialog data in a random order, neglecting rich structural information in a dataset. In this paper, we propose to use curriculum learning (CL) to better leverage both the curriculum structure and schema structure for task-oriented dialogs. Specifically, we pro...
2021.acl-short.111
10.18653/v1/2021.acl-short.111
null
2106.00291
title_judge
2021.acl-short.112
On the Generation of Medical Dialogs for COVID-19
https://aclanthology.org/2021.acl-short.112/
[ "Meng Zhou", "Zechen Li", "Bowen Tan", "Guangtao Zeng", "Wenmian Yang", "Xuehai He", "Zeqian Ju", "Subrato Chakravorty", "Shu Chen", "Xingyi Yang", "Yichen Zhang", "Qingyang Wu", "Zhou Yu", "Kun Xu", "Eric Xing", "Pengtao Xie" ]
Under the pandemic of COVID-19, people experiencing COVID19-related symptoms have a pressing need to consult doctors. Because of the shortage of medical professionals, many people cannot receive online consultations timely. To address this problem, we aim to develop a medical dialog system that can provide COVID19-rela...
2021.acl-short.112
10.18653/v1/2021.acl-short.112
null
null
null
2021.acl-short.113
Constructing Multi-Modal Dialogue Dataset by Replacing Text with Semantically Relevant Images
https://aclanthology.org/2021.acl-short.113/
[ "Nyoungwoo Lee", "Suwon Shin", "Jaegul Choo", "Ho-Jin Choi", "Sung-Hyon Myaeng" ]
In multi-modal dialogue systems, it is important to allow the use of images as part of a multi-turn conversation. Training such dialogue systems generally requires a large-scale dataset consisting of multi-turn dialogues that involve images, but such datasets rarely exist. In response, this paper proposes a 45k multi-m...
2021.acl-short.113
10.18653/v1/2021.acl-short.113
null
2107.08685
title_snapshot
2021.acl-short.114
Exposing the limits of Zero-shot Cross-lingual Hate Speech Detection
https://aclanthology.org/2021.acl-short.114/
[ "Debora Nozza" ]
Reducing and counter-acting hate speech on Social Media is a significant concern. Most of the proposed automatic methods are conducted exclusively on English and very few consistently labeled, non-English resources have been proposed. Learning to detect hate speech on English and transferring to unseen languages seems ...
2021.acl-short.114
10.18653/v1/2021.acl-short.114
null
null
null
2021.acl-short.115
BERTTune: Fine-Tuning Neural Machine Translation with BERTScore
https://aclanthology.org/2021.acl-short.115/
[ "Inigo Jauregi Unanue", "Jacob Parnell", "Massimo Piccardi" ]
Neural machine translation models are often biased toward the limited translation references seen during training. To amend this form of overfitting, in this paper we propose fine-tuning the models with a novel training objective based on the recently-proposed BERTScore evaluation metric. BERTScore is a scoring functio...
2021.acl-short.115
10.18653/v1/2021.acl-short.115
null
2106.02208
title_snapshot
2021.acl-short.116
Entity Enhancement for Implicit Discourse Relation Classification in the Biomedical Domain
https://aclanthology.org/2021.acl-short.116/
[ "Wei Shi", "Vera Demberg" ]
Implicit discourse relation classification is a challenging task, in particular when the text domain is different from the standard Penn Discourse Treebank (PDTB; Prasad et al., 2008) training corpus domain (Wall Street Journal in 1990s). We here tackle the task of implicit discourse relation classification on the biom...
2021.acl-short.116
10.18653/v1/2021.acl-short.116
null
null
null
2021.acl-short.117
Unsupervised Pronoun Resolution via Masked Noun-Phrase Prediction
https://aclanthology.org/2021.acl-short.117/
[ "Ming Shen", "Pratyay Banerjee", "Chitta Baral" ]
In this work, we propose Masked Noun-Phrase Prediction (MNPP), a pre-training strategy to tackle pronoun resolution in a fully unsupervised setting. Firstly, We evaluate our pre-trained model on various pronoun resolution datasets without any finetuning. Our method outperforms all previous unsupervised methods on all d...
2021.acl-short.117
10.18653/v1/2021.acl-short.117
null
2105.12392
title_snapshot
2021.acl-short.118
Addressing Semantic Drift in Generative Question Answering with Auxiliary Extraction
https://aclanthology.org/2021.acl-short.118/
[ "Chenliang Li", "Bin Bi", "Ming Yan", "Wei Wang", "Songfang Huang" ]
Recently, question answering (QA) based on machine reading comprehension has become popular. This work focuses on generative QA which aims to generate an abstractive answer to a given question instead of extracting an answer span from a provided passage. Generative QA often suffers from two critical problems: (1) summa...
2021.acl-short.118
10.18653/v1/2021.acl-short.118
null
null
null
2021.acl-short.119
Demoting the Lead Bias in News Summarization via Alternating Adversarial Learning
https://aclanthology.org/2021.acl-short.119/
[ "Linzi Xing", "Wen Xiao", "Giuseppe Carenini" ]
In news articles the lead bias is a common phenomenon that usually dominates the learning signals for neural extractive summarizers, severely limiting their performance on data with different or even no bias. In this paper, we introduce a novel technique to demote lead bias and make the summarizer focus more on the con...
2021.acl-short.119
10.18653/v1/2021.acl-short.119
null
2105.14241
title_snapshot
2021.acl-short.120
DuReader_robust: A Chinese Dataset Towards Evaluating Robustness and Generalization of Machine Reading Comprehension in Real-World Applications
https://aclanthology.org/2021.acl-short.120/
[ "Hongxuan Tang", "Hongyu Li", "Jing Liu", "Yu Hong", "Hua Wu", "Haifeng Wang" ]
Machine reading comprehension (MRC) is a crucial task in natural language processing and has achieved remarkable advancements. However, most of the neural MRC models are still far from robust and fail to generalize well in real-world applications. In order to comprehensively verify the robustness and generalization of ...
2021.acl-short.120
10.18653/v1/2021.acl-short.120
null
2004.11142
title_snapshot
2021.acl-short.121
Sequence to General Tree: Knowledge-Guided Geometry Word Problem Solving
https://aclanthology.org/2021.acl-short.121/
[ "Shih-hung Tsai", "Chao-Chun Liang", "Hsin-Min Wang", "Keh-Yih Su" ]
With the recent advancements in deep learning, neural solvers have gained promising results in solving math word problems. However, these SOTA solvers only generate binary expression trees that contain basic arithmetic operators and do not explicitly use the math formulas. As a result, the expression trees they produce...
2021.acl-short.121
10.18653/v1/2021.acl-short.121
null
2106.00990
title_snapshot
2021.acl-short.122
Multi-Scale Progressive Attention Network for Video Question Answering
https://aclanthology.org/2021.acl-short.122/
[ "Zhicheng Guo", "Jiaxuan Zhao", "Licheng Jiao", "Xu Liu", "Lingling Li" ]
Understanding the multi-scale visual information in a video is essential for Video Question Answering (VideoQA). Therefore, we propose a novel Multi-Scale Progressive Attention Network (MSPAN) to achieve relational reasoning between cross-scale video information. We construct clips of different lengths to represent dif...
2021.acl-short.122
10.18653/v1/2021.acl-short.122
null
null
null
2021.acl-short.123
Efficient Passage Retrieval with Hashing for Open-domain Question Answering
https://aclanthology.org/2021.acl-short.123/
[ "Ikuya Yamada", "Akari Asai", "Hannaneh Hajishirzi" ]
Most state-of-the-art open-domain question answering systems use a neural retrieval model to encode passages into continuous vectors and extract them from a knowledge source. However, such retrieval models often require large memory to run because of the massive size of their passage index. In this paper, we introduce ...
2021.acl-short.123
10.18653/v1/2021.acl-short.123
null
2106.00882
title_snapshot
2021.acl-short.124
Entity Concept-enhanced Few-shot Relation Extraction
https://aclanthology.org/2021.acl-short.124/
[ "Shan Yang", "Yongfei Zhang", "Guanglin Niu", "Qinghua Zhao", "Shiliang Pu" ]
Few-shot relation extraction (FSRE) is of great importance in long-tail distribution problem, especially in special domain with low-resource data. Most existing FSRE algorithms fail to accurately classify the relations merely based on the information of the sentences together with the recognized entity pairs, due to li...
2021.acl-short.124
10.18653/v1/2021.acl-short.124
null
2106.02401
title_snapshot
2021.acl-short.125
Improving Model Generalization: A Chinese Named Entity Recognition Case Study
https://aclanthology.org/2021.acl-short.125/
[ "Guanqing Liang", "Cane Wing-Ki Leung" ]
Generalization is an important ability that helps to ensure that a machine learning model can perform well on unseen data. In this paper, we study the effect of data bias on model generalization, using Chinese Named Entity Recognition (NER) as a case study. Specifically, we analyzed five benchmarking datasets for Chine...
2021.acl-short.125
10.18653/v1/2021.acl-short.125
null
null
null
2021.acl-short.126
Three Sentences Are All You Need: Local Path Enhanced Document Relation Extraction
https://aclanthology.org/2021.acl-short.126/
[ "Quzhe Huang", "Shengqi Zhu", "Yansong Feng", "Yuan Ye", "Yuxuan Lai", "Dongyan Zhao" ]
Document-level Relation Extraction (RE) is a more challenging task than sentence RE as it often requires reasoning over multiple sentences. Yet, human annotators usually use a small number of sentences to identify the relationship between a given entity pair. In this paper, we present an embarrassingly simple but effec...
2021.acl-short.126
10.18653/v1/2021.acl-short.126
null
2106.01793
title_snapshot
2021.acl-short.127
Unsupervised Cross-Domain Prerequisite Chain Learning using Variational Graph Autoencoders
https://aclanthology.org/2021.acl-short.127/
[ "Irene Li", "Vanessa Yan", "Tianxiao Li", "Rihao Qu", "Dragomir Radev" ]
Learning prerequisite chains is an important task for one to pick up knowledge efficiently in both known and unknown domains. For example, one may be an expert in the natural language processing (NLP) domain, but want to determine the best order in which to learn new concepts in an unfamiliar Computer Vision domain (CV...
2021.acl-short.127
10.18653/v1/2021.acl-short.127
null
2105.03505
title_snapshot
2021.acl-short.128
Attentive Multiview Text Representation for Differential Diagnosis
https://aclanthology.org/2021.acl-short.128/
[ "Hadi Amiri", "Mitra Mohtarami", "Isaac Kohane" ]
We present a text representation approach that can combine different views (representations) of the same input through effective data fusion and attention strategies for ranking purposes. We apply our model to the problem of differential diagnosis, which aims to find the most probable diseases that match with clinical ...
2021.acl-short.128
10.18653/v1/2021.acl-short.128
null
null
null
2021.acl-short.129
MedNLI Is Not Immune: Natural Language Inference Artifacts in the Clinical Domain
https://aclanthology.org/2021.acl-short.129/
[ "Christine Herlihy", "Rachel Rudinger" ]
Crowdworker-constructed natural language inference (NLI) datasets have been found to contain statistical artifacts associated with the annotation process that allow hypothesis-only classifiers to achieve better-than-random performance (CITATION). We investigate whether MedNLI, a physician-annotated dataset with premise...
2021.acl-short.129
10.18653/v1/2021.acl-short.129
null
2106.01491
title_snapshot