paper_id
stringlengths
17
19
title
stringlengths
27
144
paper_url
stringlengths
43
45
authors
listlengths
1
37
abstract
large_stringlengths
396
1.69k
anthology_id
stringlengths
17
19
doi
stringlengths
29
31
award
stringclasses
0 values
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
2021.emnlp-main.601
Learning with Different Amounts of Annotation: From Zero to Many Labels
https://aclanthology.org/2021.emnlp-main.601/
[ "Shujian Zhang", "Chengyue Gong", "Eunsol Choi" ]
Training NLP systems typically assumes access to annotated data that has a single human label per example. Given imperfect labeling from annotators and inherent ambiguity of language, we hypothesize that single label is not sufficient to learn the spectrum of language interpretation. We explore new annotation distribut...
2021.emnlp-main.601
10.18653/v1/2021.emnlp-main.601
null
2109.04408
title_snapshot
2021.emnlp-main.602
When Attention Meets Fast Recurrence: Training Language Models with Reduced Compute
https://aclanthology.org/2021.emnlp-main.602/
[ "Tao Lei" ]
Large language models have become increasingly difficult to train because of the growing computation time and cost. In this work, we present SRU++, a highly-efficient architecture that combines fast recurrence and attention for sequence modeling. SRU++ exhibits strong modeling capacity and training efficiency. On stand...
2021.emnlp-main.602
10.18653/v1/2021.emnlp-main.602
null
2102.12459
title_snapshot
2021.emnlp-main.603
Universal-KD: Attention-based Output-Grounded Intermediate Layer Knowledge Distillation
https://aclanthology.org/2021.emnlp-main.603/
[ "Yimeng Wu", "Mehdi Rezagholizadeh", "Abbas Ghaddar", "Md Akmal Haidar", "Ali Ghodsi" ]
Intermediate layer matching is shown as an effective approach for improving knowledge distillation (KD). However, this technique applies matching in the hidden spaces of two different networks (i.e. student and teacher), which lacks clear interpretability. Moreover, intermediate layer KD cannot easily deal with other p...
2021.emnlp-main.603
10.18653/v1/2021.emnlp-main.603
null
null
null
2021.emnlp-main.604
Highly Parallel Autoregressive Entity Linking with Discriminative Correction
https://aclanthology.org/2021.emnlp-main.604/
[ "Nicola De Cao", "Wilker Aziz", "Ivan Titov" ]
Generative approaches have been recently shown to be effective for both Entity Disambiguation and Entity Linking (i.e., joint mention detection and disambiguation). However, the previously proposed autoregressive formulation for EL suffers from i) high computational cost due to a complex (deep) decoder, ii) non-paralle...
2021.emnlp-main.604
10.18653/v1/2021.emnlp-main.604
null
2109.03792
title_snapshot
2021.emnlp-main.605
Word-Level Coreference Resolution
https://aclanthology.org/2021.emnlp-main.605/
[ "Vladimir Dobrovolskii" ]
Recent coreference resolution models rely heavily on span representations to find coreference links between word spans. As the number of spans is O(n^2) in the length of text and the number of potential links is O(n^4), various pruning techniques are necessary to make this approach computationally feasible. We propose ...
2021.emnlp-main.605
10.18653/v1/2021.emnlp-main.605
null
2109.04127
title_snapshot
2021.emnlp-main.606
A Secure and Efficient Federated Learning Framework for NLP
https://aclanthology.org/2021.emnlp-main.606/
[ "Chenghong Wang", "Jieren Deng", "Xianrui Meng", "Yijue Wang", "Ji Li", "Sheng Lin", "Shuo Han", "Fei Miao", "Sanguthevar Rajasekaran", "Caiwen Ding" ]
In this work, we consider the problem of designing secure and efficient federated learning (FL) frameworks for NLP. Existing solutions under this literature either consider a trusted aggregator or require heavy-weight cryptographic primitives, which makes the performance significantly degraded. Moreover, many existing ...
2021.emnlp-main.606
10.18653/v1/2021.emnlp-main.606
null
2201.11934
title_snapshot
2021.emnlp-main.607
Controllable Semantic Parsing via Retrieval Augmentation
https://aclanthology.org/2021.emnlp-main.607/
[ "Panupong Pasupat", "Yuan Zhang", "Kelvin Guu" ]
In practical applications of semantic parsing, we often want to rapidly change the behavior of the parser, such as enabling it to handle queries in a new domain, or changing its predictions on certain targeted queries. While we can introduce new training examples exhibiting the target behavior, a mechanism for enacting...
2021.emnlp-main.607
10.18653/v1/2021.emnlp-main.607
null
2110.08458
title_snapshot
2021.emnlp-main.608
Constrained Language Models Yield Few-Shot Semantic Parsers
https://aclanthology.org/2021.emnlp-main.608/
[ "Richard Shin", "Christopher Lin", "Sam Thomson", "Charles Chen", "Subhro Roy", "Emmanouil Antonios Platanios", "Adam Pauls", "Dan Klein", "Jason Eisner", "Benjamin Van Durme" ]
We explore the use of large pretrained language models as few-shot semantic parsers. The goal in semantic parsing is to generate a structured meaning representation given a natural language input. However, language models are trained to generate natural language. To bridge the gap, we use language models to paraphrase ...
2021.emnlp-main.608
10.18653/v1/2021.emnlp-main.608
null
2104.08768
title_snapshot
2021.emnlp-main.609
ExplaGraphs: An Explanation Graph Generation Task for Structured Commonsense Reasoning
https://aclanthology.org/2021.emnlp-main.609/
[ "Swarnadeep Saha", "Prateek Yadav", "Lisa Bauer", "Mohit Bansal" ]
Recent commonsense-reasoning tasks are typically discriminative in nature, where a model answers a multiple-choice question for a certain context. Discriminative tasks are limiting because they fail to adequately evaluate the model’s ability to reason and explain predictions with underlying commonsense knowledge. They ...
2021.emnlp-main.609
10.18653/v1/2021.emnlp-main.609
null
2104.07644
title_snapshot
2021.emnlp-main.610
Connect-the-Dots: Bridging Semantics between Words and Definitions via Aligning Word Sense Inventories
https://aclanthology.org/2021.emnlp-main.610/
[ "Wenlin Yao", "Xiaoman Pan", "Lifeng Jin", "Jianshu Chen", "Dian Yu", "Dong Yu" ]
Word Sense Disambiguation (WSD) aims to automatically identify the exact meaning of one word according to its context. Existing supervised models struggle to make correct predictions on rare word senses due to limited training data and can only select the best definition sentence from one predefined word sense inventor...
2021.emnlp-main.610
10.18653/v1/2021.emnlp-main.610
null
2110.14091
title_snapshot
2021.emnlp-main.611
LM-Critic: Language Models for Unsupervised Grammatical Error Correction
https://aclanthology.org/2021.emnlp-main.611/
[ "Michihiro Yasunaga", "Jure Leskovec", "Percy Liang" ]
Grammatical error correction (GEC) requires a set of labeled ungrammatical / grammatical sentence pairs for training, but obtaining such annotation can be prohibitively expensive. Recently, the Break-It-Fix-It (BIFI) framework has demonstrated strong results on learning to repair a broken program without any labeled ex...
2021.emnlp-main.611
10.18653/v1/2021.emnlp-main.611
null
2109.06822
title_snapshot
2021.emnlp-main.612
Language-agnostic Representation from Multilingual Sentence Encoders for Cross-lingual Similarity Estimation
https://aclanthology.org/2021.emnlp-main.612/
[ "Nattapong Tiyajamorn", "Tomoyuki Kajiwara", "Yuki Arase", "Makoto Onizuka" ]
We propose a method to distill a language-agnostic meaning embedding from a multilingual sentence encoder. By removing language-specific information from the original embedding, we retrieve an embedding that fully represents the sentence’s meaning. The proposed method relies only on parallel corpora without any human a...
2021.emnlp-main.612
10.18653/v1/2021.emnlp-main.612
null
null
null
2021.emnlp-main.613
Classifying Dyads for Militarized Conflict Analysis
https://aclanthology.org/2021.emnlp-main.613/
[ "Niklas Stoehr", "Lucas Torroba Hennigen", "Samin Ahbab", "Robert West", "Ryan Cotterell" ]
Understanding the origins of militarized conflict is a complex, yet important undertaking. Existing research seeks to build this understanding by considering bi-lateral relationships between entity pairs (dyadic causes) and multi-lateral relationships among multiple entities (systemic causes). The aim of this work is t...
2021.emnlp-main.613
10.18653/v1/2021.emnlp-main.613
null
2109.12860
title_snapshot
2021.emnlp-main.614
Point-of-Interest Type Prediction using Text and Images
https://aclanthology.org/2021.emnlp-main.614/
[ "Danae Sánchez Villegas", "Nikolaos Aletras" ]
Point-of-interest (POI) type prediction is the task of inferring the type of a place from where a social media post was shared. Inferring a POI’s type is useful for studies in computational social science including sociolinguistics, geosemiotics, and cultural geography, and has applications in geosocial networking tech...
2021.emnlp-main.614
10.18653/v1/2021.emnlp-main.614
null
2109.00602
title_snapshot
2021.emnlp-main.615
Come hither or go away? Recognising pre-electoral coalition signals in the news
https://aclanthology.org/2021.emnlp-main.615/
[ "Ines Rehbein", "Simone Paolo Ponzetto", "Anna Adendorf", "Oke Bahnsen", "Lukas Stoetzer", "Heiner Stuckenschmidt" ]
In this paper, we introduce the task of political coalition signal prediction from text, that is, the task of recognizing from the news coverage leading up to an election the (un)willingness of political parties to form a government coalition. We decompose our problem into two related, but distinct tasks: (i) predictin...
2021.emnlp-main.615
10.18653/v1/2021.emnlp-main.615
null
null
null
2021.emnlp-main.616
#HowYouTagTweets: Learning User Hashtagging Preferences via Personalized Topic Attention
https://aclanthology.org/2021.emnlp-main.616/
[ "Yuji Zhang", "Yubo Zhang", "Chunpu Xu", "Jing Li", "Ziyan Jiang", "Baolin Peng" ]
Millions of hashtags are created on social media every day to cross-refer messages concerning similar topics. To help people find the topics they want to discuss, this paper characterizes a user’s hashtagging preferences via predicting how likely they will post with a hashtag. It is hypothesized that one’s interests in...
2021.emnlp-main.616
10.18653/v1/2021.emnlp-main.616
null
null
null
2021.emnlp-main.617
Learning Neural Templates for Recommender Dialogue System
https://aclanthology.org/2021.emnlp-main.617/
[ "Zujie Liang", "Huang Hu", "Can Xu", "Jian Miao", "Yingying He", "Yining Chen", "Xiubo Geng", "Fan Liang", "Daxin Jiang" ]
The task of Conversational Recommendation System (CRS), i.e., recommender dialog system, aims to recommend precise items to users through natural language interactions. Though recent end-to-end neural models have shown promising progress on this task, two key challenges still remain. First, the recommended items cannot...
2021.emnlp-main.617
10.18653/v1/2021.emnlp-main.617
null
2109.12302
title_snapshot
2021.emnlp-main.618
Proxy Indicators for the Quality of Open-domain Dialogues
https://aclanthology.org/2021.emnlp-main.618/
[ "Rostislav Nedelchev", "Jens Lehmann", "Ricardo Usbeck" ]
The automatic evaluation of open-domain dialogues remains a largely unsolved challenge. Despite the abundance of work done in the field, human judges have to evaluate dialogues’ quality. As a consequence, performing such evaluations at scale is usually expensive. This work investigates using a deep-learning model train...
2021.emnlp-main.618
10.18653/v1/2021.emnlp-main.618
null
null
null
2021.emnlp-main.619
Q^{2}: Evaluating Factual Consistency in Knowledge-Grounded Dialogues via Question Generation and Question Answering
https://aclanthology.org/2021.emnlp-main.619/
[ "Or Honovich", "Leshem Choshen", "Roee Aharoni", "Ella Neeman", "Idan Szpektor", "Omri Abend" ]
Neural knowledge-grounded generative models for dialogue often produce content that is factually inconsistent with the knowledge they rely on, making them unreliable and limiting their applicability. Inspired by recent work on evaluating factual consistency in abstractive summarization, we propose an automatic evaluati...
2021.emnlp-main.619
10.18653/v1/2021.emnlp-main.619
null
2104.08202
title_snapshot
2021.emnlp-main.620
Knowledge-Aware Graph-Enhanced GPT-2 for Dialogue State Tracking
https://aclanthology.org/2021.emnlp-main.620/
[ "Weizhe Lin", "Bo-Hsiang Tseng", "Bill Byrne" ]
Dialogue State Tracking is central to multi-domain task-oriented dialogue systems, responsible for extracting information from user utterances. We present a novel hybrid architecture that augments GPT-2 with representations derived from Graph Attention Networks in such a way to allow causal, sequential prediction of sl...
2021.emnlp-main.620
10.18653/v1/2021.emnlp-main.620
null
2104.04466
title_snapshot
2021.emnlp-main.621
A Collaborative Multi-agent Reinforcement Learning Framework for Dialog Action Decomposition
https://aclanthology.org/2021.emnlp-main.621/
[ "Huimin Wang", "Kam-Fai Wong" ]
Most reinforcement learning methods for dialog policy learning train a centralized agent that selects a predefined joint action concatenating domain name, intent type, and slot name. The centralized dialog agent suffers from a great many user-agent interaction requirements due to the large action space. Besides, design...
2021.emnlp-main.621
10.18653/v1/2021.emnlp-main.621
null
null
null
2021.emnlp-main.622
Zero-Shot Dialogue State Tracking via Cross-Task Transfer
https://aclanthology.org/2021.emnlp-main.622/
[ "Zhaojiang Lin", "Bing Liu", "Andrea Madotto", "Seungwhan Moon", "Zhenpeng Zhou", "Paul Crook", "Zhiguang Wang", "Zhou Yu", "Eunjoon Cho", "Rajen Subba", "Pascale Fung" ]
Zero-shot transfer learning for dialogue state tracking (DST) enables us to handle a variety of task-oriented dialogue domains without the expense of collecting in-domain data. In this work, we propose to transfer the cross-task knowledge from general question answering (QA) corpora for the zero-shot DST task. Specific...
2021.emnlp-main.622
10.18653/v1/2021.emnlp-main.622
null
2109.04655
title_snapshot
2021.emnlp-main.623
Uncertainty Measures in Neural Belief Tracking and the Effects on Dialogue Policy Performance
https://aclanthology.org/2021.emnlp-main.623/
[ "Carel van Niekerk", "Andrey Malinin", "Christian Geishauser", "Michael Heck", "Hsien-chin Lin", "Nurul Lubis", "Shutong Feng", "Milica Gasic" ]
The ability to identify and resolve uncertainty is crucial for the robustness of a dialogue system. Indeed, this has been confirmed empirically on systems that utilise Bayesian approaches to dialogue belief tracking. However, such systems consider only confidence estimates and have difficulty scaling to more complex se...
2021.emnlp-main.623
10.18653/v1/2021.emnlp-main.623
null
2109.04349
title_snapshot
2021.emnlp-main.624
Dynamic Forecasting of Conversation Derailment
https://aclanthology.org/2021.emnlp-main.624/
[ "Yova Kementchedjhieva", "Anders Søgaard" ]
Online conversations can sometimes take a turn for the worse, either due to systematic cultural differences, accidental misunderstandings, or mere malice. Automatically forecasting derailment in public online conversations provides an opportunity to take early action to moderate it. Previous work in this space is limit...
2021.emnlp-main.624
10.18653/v1/2021.emnlp-main.624
null
2110.05111
title_snapshot
2021.emnlp-main.625
A Semantic Filter Based on Relations for Knowledge Graph Completion
https://aclanthology.org/2021.emnlp-main.625/
[ "Zongwei Liang", "Junan Yang", "Hui Liu", "Keju Huang" ]
Knowledge graph embedding, representing entities and relations in the knowledge graphs with high-dimensional vectors, has made significant progress in link prediction. More researchers have explored the representational capabilities of models in recent years. That is, they investigate better representational models to ...
2021.emnlp-main.625
10.18653/v1/2021.emnlp-main.625
null
null
null
2021.emnlp-main.626
AdapterDrop: On the Efficiency of Adapters in Transformers
https://aclanthology.org/2021.emnlp-main.626/
[ "Andreas Rücklé", "Gregor Geigle", "Max Glockner", "Tilman Beck", "Jonas Pfeiffer", "Nils Reimers", "Iryna Gurevych" ]
Transformer models are expensive to fine-tune, slow for inference, and have large storage requirements. Recent approaches tackle these shortcomings by training smaller models, dynamically reducing the model size, and by training light-weight adapters. In this paper, we propose AdapterDrop, removing adapters from lower ...
2021.emnlp-main.626
10.18653/v1/2021.emnlp-main.626
null
2010.11918
title_snapshot
2021.emnlp-main.627
Understanding and Overcoming the Challenges of Efficient Transformer Quantization
https://aclanthology.org/2021.emnlp-main.627/
[ "Yelysei Bondarenko", "Markus Nagel", "Tijmen Blankevoort" ]
Transformer-based architectures have become the de-facto standard models for a wide range of Natural Language Processing tasks. However, their memory footprint and high latency are prohibitive for efficient deployment and inference on resource-limited devices. In this work, we explore quantization for transformers. We ...
2021.emnlp-main.627
10.18653/v1/2021.emnlp-main.627
null
2109.12948
title_snapshot
2021.emnlp-main.628
CAPE: Context-Aware Private Embeddings for Private Language Learning
https://aclanthology.org/2021.emnlp-main.628/
[ "Richard Plant", "Dimitra Gkatzia", "Valerio Giuffrida" ]
Neural language models have contributed to state-of-the-art results in a number of downstream applications including sentiment analysis, intent classification and others. However, obtaining text representations or embeddings using these models risks encoding personally identifiable information learned from language and...
2021.emnlp-main.628
10.18653/v1/2021.emnlp-main.628
null
2108.12318
title_snapshot
2021.emnlp-main.629
Text Detoxification using Large Pre-trained Neural Models
https://aclanthology.org/2021.emnlp-main.629/
[ "David Dale", "Anton Voronov", "Daryna Dementieva", "Varvara Logacheva", "Olga Kozlova", "Nikita Semenov", "Alexander Panchenko" ]
We present two novel unsupervised methods for eliminating toxicity in text. Our first method combines two recent ideas: (1) guidance of the generation process with small style-conditional language models and (2) use of paraphrasing models to perform style transfer. We use a well-performing paraphraser guided by style-t...
2021.emnlp-main.629
10.18653/v1/2021.emnlp-main.629
null
2109.08914
title_snapshot
2021.emnlp-main.630
Document-Level Text Simplification: Dataset, Criteria and Baseline
https://aclanthology.org/2021.emnlp-main.630/
[ "Renliang Sun", "Hanqi Jin", "Xiaojun Wan" ]
Text simplification is a valuable technique. However, current research is limited to sentence simplification. In this paper, we define and investigate a new task of document-level text simplification, which aims to simplify a document consisting of multiple sentences. Based on Wikipedia dumps, we first construct a larg...
2021.emnlp-main.630
10.18653/v1/2021.emnlp-main.630
null
2110.05071
title_snapshot
2021.emnlp-main.631
A Bag of Tricks for Dialogue Summarization
https://aclanthology.org/2021.emnlp-main.631/
[ "Muhammad Khalifa", "Miguel Ballesteros", "Kathleen McKeown" ]
Dialogue summarization comes with its own peculiar challenges as opposed to news or scientific articles summarization. In this work, we explore four different challenges of the task: handling and differentiating parts of the dialogue belonging to multiple speakers, negation understanding, reasoning about the situation,...
2021.emnlp-main.631
10.18653/v1/2021.emnlp-main.631
null
2109.08232
title_snapshot
2021.emnlp-main.632
Paraphrasing Compound Nominalizations
https://aclanthology.org/2021.emnlp-main.632/
[ "John Lee", "Ho Hung Lim", "Carol Webster" ]
A nominalization uses a deverbal noun to describe an event associated with its underlying verb. Commonly found in academic and formal texts, nominalizations can be difficult to interpret because of ambiguous semantic relations between the deverbal noun and its arguments. Our goal is to interpret nominalizations by gene...
2021.emnlp-main.632
10.18653/v1/2021.emnlp-main.632
null
null
null
2021.emnlp-main.633
Data-QuestEval: A Referenceless Metric for Data-to-Text Semantic Evaluation
https://aclanthology.org/2021.emnlp-main.633/
[ "Clement Rebuffel", "Thomas Scialom", "Laure Soulier", "Benjamin Piwowarski", "Sylvain Lamprier", "Jacopo Staiano", "Geoffrey Scoutheeten", "Patrick Gallinari" ]
QuestEval is a reference-less metric used in text-to-text tasks, that compares the generated summaries directly to the source text, by automatically asking and answering questions. Its adaptation to Data-to-Text tasks is not straightforward, as it requires multimodal Question Generation and Answering systems on the con...
2021.emnlp-main.633
10.18653/v1/2021.emnlp-main.633
null
2104.07555
title_snapshot
2021.emnlp-main.634
Low-Rank Subspaces for Unsupervised Entity Linking
https://aclanthology.org/2021.emnlp-main.634/
[ "Akhil Arora", "Alberto Garcia-Duran", "Robert West" ]
Entity linking is an important problem with many applications. Most previous solutions were designed for settings where annotated training data is available, which is, however, not the case in numerous domains. We propose a light-weight and scalable entity linking method, Eigenthemes, that relies solely on the availabi...
2021.emnlp-main.634
10.18653/v1/2021.emnlp-main.634
null
2104.08737
title_snapshot
2021.emnlp-main.635
TDEER: An Efficient Translating Decoding Schema for Joint Extraction of Entities and Relations
https://aclanthology.org/2021.emnlp-main.635/
[ "Xianming Li", "Xiaotian Luo", "Chenghao Dong", "Daichuan Yang", "Beidi Luan", "Zhen He" ]
Joint extraction of entities and relations from unstructured texts to form factual triples is a fundamental task of constructing a Knowledge Base (KB). A common method is to decode triples by predicting entity pairs to obtain the corresponding relation. However, it is still challenging to handle this task efficiently, ...
2021.emnlp-main.635
10.18653/v1/2021.emnlp-main.635
null
null
null
2021.emnlp-main.636
Extracting Event Temporal Relations via Hyperbolic Geometry
https://aclanthology.org/2021.emnlp-main.636/
[ "Xingwei Tan", "Gabriele Pergola", "Yulan He" ]
Detecting events and their evolution through time is a crucial task in natural language understanding. Recent neural approaches to event temporal relation extraction typically map events to embeddings in the Euclidean space and train a classifier to detect temporal relations between event pairs. However, embeddings in ...
2021.emnlp-main.636
10.18653/v1/2021.emnlp-main.636
null
2109.05527
title_snapshot
2021.emnlp-main.637
Honey or Poison? Solving the Trigger Curse in Few-shot Event Detection via Causal Intervention
https://aclanthology.org/2021.emnlp-main.637/
[ "Jiawei Chen", "Hongyu Lin", "Xianpei Han", "Le Sun" ]
Event detection has long been troubled by the trigger curse: overfitting the trigger will harm the generalization ability while underfitting it will hurt the detection performance. This problem is even more severe in few-shot scenario. In this paper, we identify and solve the trigger curse problem in few-shot event det...
2021.emnlp-main.637
10.18653/v1/2021.emnlp-main.637
null
2109.05747
title_snapshot
2021.emnlp-main.638
Back to the Basics: A Quantitative Analysis of Statistical and Graph-Based Term Weighting Schemes for Keyword Extraction
https://aclanthology.org/2021.emnlp-main.638/
[ "Asahi Ushio", "Federico Liberatore", "Jose Camacho-Collados" ]
Term weighting schemes are widely used in Natural Language Processing and Information Retrieval. In particular, term weighting is the basis for keyword extraction. However, there are relatively few evaluation studies that shed light about the strengths and shortcomings of each weighting scheme. In fact, in most cases r...
2021.emnlp-main.638
10.18653/v1/2021.emnlp-main.638
null
2104.08028
title_snapshot
2021.emnlp-main.639
Time-dependent Entity Embedding is not All You Need: A Re-evaluation of Temporal Knowledge Graph Completion Models under a Unified Framework
https://aclanthology.org/2021.emnlp-main.639/
[ "Zhen Han", "Gengyuan Zhang", "Yunpu Ma", "Volker Tresp" ]
Various temporal knowledge graph (KG) completion models have been proposed in the recent literature. The models usually contain two parts, a temporal embedding layer and a score function derived from existing static KG modeling approaches. Since the approaches differ along several dimensions, including different score ...
2021.emnlp-main.639
10.18653/v1/2021.emnlp-main.639
null
null
null
2021.emnlp-main.640
Matching-oriented Embedding Quantization For Ad-hoc Retrieval
https://aclanthology.org/2021.emnlp-main.640/
[ "Shitao Xiao", "Zheng Liu", "Yingxia Shao", "Defu Lian", "Xing Xie" ]
Product quantization (PQ) is a widely used technique for ad-hoc retrieval. Recent studies propose supervised PQ, where the embedding and quantization models can be jointly trained with supervised learning. However, there is a lack of appropriate formulation of the joint training objective; thus, the improvements over p...
2021.emnlp-main.640
10.18653/v1/2021.emnlp-main.640
null
2104.07858
title_judge
2021.emnlp-main.641
Efficient Mind-Map Generation via Sequence-to-Graph and Reinforced Graph Refinement
https://aclanthology.org/2021.emnlp-main.641/
[ "Mengting Hu", "Honglei Guo", "Shiwan Zhao", "Hang Gao", "Zhong Su" ]
A mind-map is a diagram that represents the central concept and key ideas in a hierarchical way. Converting plain text into a mind-map will reveal its key semantic structure and be easier to understand. Given a document, the existing automatic mind-map generation method extracts the relationships of every sentence pair...
2021.emnlp-main.641
10.18653/v1/2021.emnlp-main.641
null
2109.02457
title_snapshot
2021.emnlp-main.642
Deep Attention Diffusion Graph Neural Networks for Text Classification
https://aclanthology.org/2021.emnlp-main.642/
[ "Yonghao Liu", "Renchu Guan", "Fausto Giunchiglia", "Yanchun Liang", "Xiaoyue Feng" ]
Text classification is a fundamental task with broad applications in natural language processing. Recently, graph neural networks (GNNs) have attracted much attention due to their powerful representation ability. However, most existing methods for text classification based on GNNs consider only one-hop neighborhoods an...
2021.emnlp-main.642
10.18653/v1/2021.emnlp-main.642
null
null
null
2021.emnlp-main.643
Balancing Methods for Multi-label Text Classification with Long-Tailed Class Distribution
https://aclanthology.org/2021.emnlp-main.643/
[ "Yi Huang", "Buse Giledereli", "Abdullatif Köksal", "Arzucan Özgür", "Elif Ozkirimli" ]
Multi-label text classification is a challenging task because it requires capturing label dependencies. It becomes even more challenging when class distribution is long-tailed. Resampling and re-weighting are common approaches used for addressing the class imbalance problem, however, they are not effective when there i...
2021.emnlp-main.643
10.18653/v1/2021.emnlp-main.643
null
2109.04712
title_snapshot
2021.emnlp-main.644
Bayesian Topic Regression for Causal Inference
https://aclanthology.org/2021.emnlp-main.644/
[ "Maximilian Ahrens", "Julian Ashwin", "Jan-Peter Calliess", "Vu Nguyen" ]
Causal inference using observational text data is becoming increasingly popular in many research areas. This paper presents the Bayesian Topic Regression (BTR) model that uses both text and numerical information to model an outcome variable. It allows estimation of both discrete and continuous treatment effects. Furthe...
2021.emnlp-main.644
10.18653/v1/2021.emnlp-main.644
null
2109.05317
title_snapshot
2021.emnlp-main.645
Enjoy the Salience: Towards Better Transformer-based Faithful Explanations with Word Salience
https://aclanthology.org/2021.emnlp-main.645/
[ "George Chrysostomou", "Nikolaos Aletras" ]
Pretrained transformer-based models such as BERT have demonstrated state-of-the-art predictive performance when adapted into a range of natural language processing tasks. An open problem is how to improve the faithfulness of explanations (rationales) for the predictions of these models. In this paper, we hypothesize th...
2021.emnlp-main.645
10.18653/v1/2021.emnlp-main.645
null
2108.13759
title_snapshot
2021.emnlp-main.646
What’s in Your Head? Emergent Behaviour in Multi-Task Transformer Models
https://aclanthology.org/2021.emnlp-main.646/
[ "Mor Geva", "Uri Katz", "Aviv Ben-Arie", "Jonathan Berant" ]
The primary paradigm for multi-task training in natural language processing is to represent the input with a shared pre-trained language model, and add a small, thin network (head) per task. Given an input, a target head is the head that is selected for outputting the final prediction. In this work, we examine the beha...
2021.emnlp-main.646
10.18653/v1/2021.emnlp-main.646
null
2104.06129
title_snapshot
2021.emnlp-main.647
Don’t Search for a Search Method — Simple Heuristics Suffice for Adversarial Text Attacks
https://aclanthology.org/2021.emnlp-main.647/
[ "Nathaniel Berger", "Stefan Riezler", "Sebastian Ebert", "Artem Sokolov" ]
Recently more attention has been given to adversarial attacks on neural networks for natural language processing (NLP). A central research topic has been the investigation of search algorithms and search constraints, accompanied by benchmark algorithms and tasks. We implement an algorithm inspired by zeroth order optim...
2021.emnlp-main.647
10.18653/v1/2021.emnlp-main.647
null
2109.07926
title_snapshot
2021.emnlp-main.648
Adversarial Attacks on Knowledge Graph Embeddings via Instance Attribution Methods
https://aclanthology.org/2021.emnlp-main.648/
[ "Peru Bhardwaj", "John Kelleher", "Luca Costabello", "Declan O’Sullivan" ]
Despite the widespread use of Knowledge Graph Embeddings (KGE), little is known about the security vulnerabilities that might disrupt their intended behaviour. We study data poisoning attacks against KGE models for link prediction. These attacks craft adversarial additions or deletions at training time to cause model f...
2021.emnlp-main.648
10.18653/v1/2021.emnlp-main.648
null
2111.03120
title_snapshot
2021.emnlp-main.649
Locke’s Holiday: Belief Bias in Machine Reading
https://aclanthology.org/2021.emnlp-main.649/
[ "Anders Søgaard" ]
I highlight a simple failure mode of state-of-the-art machine reading systems: when contexts do not align with commonly shared beliefs. For example, machine reading systems fail to answer What did Elizabeth want? correctly in the context of ‘My kingdom for a cough drop, cried Queen Elizabeth.’ Biased by co-occurrence s...
2021.emnlp-main.649
10.18653/v1/2021.emnlp-main.649
null
null
null
2021.emnlp-main.650
Sequence Length is a Domain: Length-based Overfitting in Transformer Models
https://aclanthology.org/2021.emnlp-main.650/
[ "Dusan Varis", "Ondřej Bojar" ]
Transformer-based sequence-to-sequence architectures, while achieving state-of-the-art results on a large number of NLP tasks, can still suffer from overfitting during training. In practice, this is usually countered either by applying regularization methods (e.g. dropout, L2-regularization) or by providing huge amount...
2021.emnlp-main.650
10.18653/v1/2021.emnlp-main.650
null
2109.07276
title_snapshot
2021.emnlp-main.651
Contrasting Human- and Machine-Generated Word-Level Adversarial Examples for Text Classification
https://aclanthology.org/2021.emnlp-main.651/
[ "Maximilian Mozes", "Max Bartolo", "Pontus Stenetorp", "Bennett Kleinberg", "Lewis Griffin" ]
Research shows that natural language processing models are generally considered to be vulnerable to adversarial attacks; but recent work has drawn attention to the issue of validating these adversarial inputs against certain criteria (e.g., the preservation of semantics and grammaticality). Enforcing constraints to uph...
2021.emnlp-main.651
10.18653/v1/2021.emnlp-main.651
null
2109.04385
title_snapshot
2021.emnlp-main.652
Is Information Density Uniform in Task-Oriented Dialogues?
https://aclanthology.org/2021.emnlp-main.652/
[ "Mario Giulianelli", "Arabella Sinclair", "Raquel Fernández" ]
The Uniform Information Density principle states that speakers plan their utterances to reduce fluctuations in the density of the information transmitted. In this paper, we test whether, and within which contextual units this principle holds in task-oriented dialogues. We show that there is evidence supporting the prin...
2021.emnlp-main.652
10.18653/v1/2021.emnlp-main.652
null
null
null
2021.emnlp-main.653
On Homophony and Rényi Entropy
https://aclanthology.org/2021.emnlp-main.653/
[ "Tiago Pimentel", "Clara Meister", "Simone Teufel", "Ryan Cotterell" ]
Homophony’s widespread presence in natural languages is a controversial topic. Recent theories of language optimality have tried to justify its prevalence, despite its negative effects on cognitive processing time, e.g., Piantadosi et al. (2012) argued homophony enables the reuse of efficient wordforms and is thus bene...
2021.emnlp-main.653
10.18653/v1/2021.emnlp-main.653
null
2109.13766
title_snapshot
2021.emnlp-main.654
Synthetic Textual Features for the Large-Scale Detection of Basic-level Categories in English and Mandarin
https://aclanthology.org/2021.emnlp-main.654/
[ "Yiwen Chen", "Simone Teufel" ]
Basic-level categories (BLC) are an important psycholinguistic concept introduced by Rosch et al. (1976); they are defined as the most inclusive categories for which a concrete mental image of the category as a whole can be formed, and also as those categories which are acquired early in life. Rosch’s original algorith...
2021.emnlp-main.654
10.18653/v1/2021.emnlp-main.654
null
null
null
2021.emnlp-main.655
TimeTraveler: Reinforcement Learning for Temporal Knowledge Graph Forecasting
https://aclanthology.org/2021.emnlp-main.655/
[ "Haohai Sun", "Jialun Zhong", "Yunpu Ma", "Zhen Han", "Kun He" ]
Temporal knowledge graph (TKG) reasoning is a crucial task that has gained increasing research interest in recent years. Most existing methods focus on reasoning at past timestamps to complete the missing facts, and there are only a few works of reasoning on known TKGs to forecast future facts. Compared with the comple...
2021.emnlp-main.655
10.18653/v1/2021.emnlp-main.655
null
2109.04101
title_snapshot
2021.emnlp-main.656
Code-switched inspired losses for spoken dialog representations
https://aclanthology.org/2021.emnlp-main.656/
[ "Pierre Colombo", "Emile Chapuis", "Matthieu Labeau", "Chloé Clavel" ]
Spoken dialogue systems need to be able to handle both multiple languages and multilinguality inside a conversation (e.g in case of code-switching). In this work, we introduce new pretraining losses tailored to learn generic multilingual spoken dialogue representations. The goal of these losses is to expose the model t...
2021.emnlp-main.656
10.18653/v1/2021.emnlp-main.656
null
2108.12465
title_judge
2021.emnlp-main.657
BiQUE: Biquaternionic Embeddings of Knowledge Graphs
https://aclanthology.org/2021.emnlp-main.657/
[ "Jia Guo", "Stanley Kok" ]
Knowledge graph embeddings (KGEs) compactly encode multi-relational knowledge graphs (KGs). Existing KGE models rely on geometric operations to model relational patterns. Euclidean (circular) rotation is useful for modeling patterns such as symmetry, but cannot represent hierarchical semantics. In contrast, hyperbolic ...
2021.emnlp-main.657
10.18653/v1/2021.emnlp-main.657
null
2109.14401
title_snapshot
2021.emnlp-main.658
Learning Neural Ordinary Equations for Forecasting Future Links on Temporal Knowledge Graphs
https://aclanthology.org/2021.emnlp-main.658/
[ "Zhen Han", "Zifeng Ding", "Yunpu Ma", "Yujia Gu", "Volker Tresp" ]
There has been an increasing interest in inferring future links on temporal knowledge graphs (KG). While links on temporal KGs vary continuously over time, the existing approaches model the temporal KGs in discrete state spaces. To this end, we propose a novel continuum model by extending the idea of neural ordinary di...
2021.emnlp-main.658
10.18653/v1/2021.emnlp-main.658
null
null
null
2021.emnlp-main.659
RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models
https://aclanthology.org/2021.emnlp-main.659/
[ "Wenkai Yang", "Yankai Lin", "Peng Li", "Jie Zhou", "Xu Sun" ]
Backdoor attacks, which maliciously control a well-trained model’s outputs of the instances with specific triggers, are recently shown to be serious threats to the safety of reusing deep neural networks (DNNs). In this work, we propose an efficient online defense mechanism based on robustness-aware perturbations. Speci...
2021.emnlp-main.659
10.18653/v1/2021.emnlp-main.659
null
2110.07831
title_snapshot
2021.emnlp-main.660
FAME: Feature-Based Adversarial Meta-Embeddings for Robust Input Representations
https://aclanthology.org/2021.emnlp-main.660/
[ "Lukas Lange", "Heike Adel", "Jannik Strötgen", "Dietrich Klakow" ]
Combining several embeddings typically improves performance in downstream tasks as different embeddings encode different information. It has been shown that even models using embeddings from transformers still benefit from the inclusion of standard word embeddings. However, the combination of embeddings of different ty...
2021.emnlp-main.660
10.18653/v1/2021.emnlp-main.660
null
2010.12305
title_snapshot
2021.emnlp-main.661
A Strong Baseline for Query Efficient Attacks in a Black Box Setting
https://aclanthology.org/2021.emnlp-main.661/
[ "Rishabh Maheshwary", "Saket Maheshwary", "Vikram Pudi" ]
Existing black box search methods have achieved high success rate in generating adversarial attacks against NLP models. However, such search methods are inefficient as they do not consider the amount of queries required to generate adversarial attacks. Also, prior attacks do not maintain a consistent search space while...
2021.emnlp-main.661
10.18653/v1/2021.emnlp-main.661
null
2109.04775
title_snapshot
2021.emnlp-main.662
Machine Translation Decoding beyond Beam Search
https://aclanthology.org/2021.emnlp-main.662/
[ "Rémi Leblond", "Jean-Baptiste Alayrac", "Laurent Sifre", "Miruna Pislar", "Lespiau Jean-Baptiste", "Ioannis Antonoglou", "Karen Simonyan", "Oriol Vinyals" ]
Beam search is the go-to method for decoding auto-regressive machine translation models. While it yields consistent improvements in terms of BLEU, it is only concerned with finding outputs with high model likelihood, and is thus agnostic to whatever end metric or score practitioners care about. Our aim is to establish ...
2021.emnlp-main.662
10.18653/v1/2021.emnlp-main.662
null
2104.05336
title_snapshot
2021.emnlp-main.663
Document Graph for Neural Machine Translation
https://aclanthology.org/2021.emnlp-main.663/
[ "Mingzhou Xu", "Liangyou Li", "Derek F. Wong", "Qun Liu", "Lidia S. Chao" ]
Previous works have shown that contextual information can improve the performance of neural machine translation (NMT). However, most existing document-level NMT methods failed to leverage contexts beyond a few set of previous sentences. How to make use of the whole document as global contexts is still a challenge. To a...
2021.emnlp-main.663
10.18653/v1/2021.emnlp-main.663
null
2012.03477
title_snapshot
2021.emnlp-main.664
An Empirical Investigation of Word Alignment Supervision for Zero-Shot Multilingual Neural Machine Translation
https://aclanthology.org/2021.emnlp-main.664/
[ "Alessandro Raganato", "Raúl Vázquez", "Mathias Creutz", "Jörg Tiedemann" ]
Zero-shot translations is a fascinating feature of Multilingual Neural Machine Translation (MNMT) systems. These MNMT models are usually trained on English-centric data, i.e. English either as the source or target language, and with a language label prepended to the input indicating the target language. However, recent...
2021.emnlp-main.664
10.18653/v1/2021.emnlp-main.664
null
null
null
2021.emnlp-main.665
Graph Algorithms for Multiparallel Word Alignment
https://aclanthology.org/2021.emnlp-main.665/
[ "Ayyoob Imani", "Masoud Jalili Sabet", "Lutfi Kerem Senel", "Philipp Dufter", "François Yvon", "Hinrich Schütze" ]
With the advent of end-to-end deep learning approaches in machine translation, interest in word alignments initially decreased; however, they have again become a focus of research more recently. Alignments are useful for typological research, transferring formatting like markup to translated texts, and can be used in t...
2021.emnlp-main.665
10.18653/v1/2021.emnlp-main.665
null
2109.06283
title_snapshot
2021.emnlp-main.666
Improving the Quality Trade-Off for Neural Machine Translation Multi-Domain Adaptation
https://aclanthology.org/2021.emnlp-main.666/
[ "Eva Hasler", "Tobias Domhan", "Jonay Trenous", "Ke Tran", "Bill Byrne", "Felix Hieber" ]
Building neural machine translation systems to perform well on a specific target domain is a well-studied problem. Optimizing system performance for multiple, diverse target domains however remains a challenge. We study this problem in an adaptation setting where the goal is to preserve the existing system quality whil...
2021.emnlp-main.666
10.18653/v1/2021.emnlp-main.666
null
null
null
2021.emnlp-main.667
Language Modeling, Lexical Translation, Reordering: The Training Process of NMT through the Lens of Classical SMT
https://aclanthology.org/2021.emnlp-main.667/
[ "Elena Voita", "Rico Sennrich", "Ivan Titov" ]
Differently from the traditional statistical MT that decomposes the translation task into distinct separately learned components, neural machine translation uses a single neural network to model the entire translation process. Despite neural machine translation being de-facto standard, it is still not clear how NMT mod...
2021.emnlp-main.667
10.18653/v1/2021.emnlp-main.667
null
2109.01396
title_snapshot
2021.emnlp-main.668
Effective Fine-Tuning Methods for Cross-lingual Adaptation
https://aclanthology.org/2021.emnlp-main.668/
[ "Tao Yu", "Shafiq Joty" ]
Large scale multilingual pre-trained language models have shown promising results in zero- and few-shot cross-lingual tasks. However, recent studies have shown their lack of generalizability when the languages are structurally dissimilar. In this work, we propose a novel fine-tuning method based on co-training that aim...
2021.emnlp-main.668
10.18653/v1/2021.emnlp-main.668
null
null
null
2021.emnlp-main.669
Rethinking Data Augmentation for Low-Resource Neural Machine Translation: A Multi-Task Learning Approach
https://aclanthology.org/2021.emnlp-main.669/
[ "Víctor M. Sánchez-Cartagena", "Miquel Esplà-Gomis", "Juan Antonio Pérez-Ortiz", "Felipe Sánchez-Martínez" ]
In the context of neural machine translation, data augmentation (DA) techniques may be used for generating additional training samples when the available parallel data are scarce. Many DA approaches aim at expanding the support of the empirical data distribution by generating new sentence pairs that contain infrequent ...
2021.emnlp-main.669
10.18653/v1/2021.emnlp-main.669
null
2109.03645
title_snapshot
2021.emnlp-main.670
Wino-X: Multilingual Winograd Schemas for Commonsense Reasoning and Coreference Resolution
https://aclanthology.org/2021.emnlp-main.670/
[ "Denis Emelin", "Rico Sennrich" ]
Winograd schemas are a well-established tool for evaluating coreference resolution (CoR) and commonsense reasoning (CSR) capabilities of computational models. So far, schemas remained largely confined to English, limiting their utility in multilingual settings. This work presents Wino-X, a parallel dataset of German, F...
2021.emnlp-main.670
10.18653/v1/2021.emnlp-main.670
null
null
null
2021.emnlp-main.671
One Source, Two Targets: Challenges and Rewards of Dual Decoding
https://aclanthology.org/2021.emnlp-main.671/
[ "Jitao Xu", "François Yvon" ]
Machine translation is generally understood as generating one target text from an input source document. In this paper, we consider a stronger requirement: to jointly generate two texts so that each output side effectively depends on the other. As we discuss, such a device serves several practical purposes, from multi-...
2021.emnlp-main.671
10.18653/v1/2021.emnlp-main.671
null
2109.10197
title_snapshot
2021.emnlp-main.672
Discrete and Soft Prompting for Multilingual Models
https://aclanthology.org/2021.emnlp-main.672/
[ "Mengjie Zhao", "Hinrich Schütze" ]
It has been shown for English that discrete and soft prompting perform strongly in few-shot learning with pretrained language models (PLMs). In this paper, we show that discrete and soft prompting perform better than finetuning in multilingual cases: Crosslingual transfer and in-language training of multilingual natura...
2021.emnlp-main.672
10.18653/v1/2021.emnlp-main.672
null
2109.03630
title_snapshot
2021.emnlp-main.673
Vision Matters When It Should: Sanity Checking Multimodal Machine Translation Models
https://aclanthology.org/2021.emnlp-main.673/
[ "Jiaoda Li", "Duygu Ataman", "Rico Sennrich" ]
Multimodal machine translation (MMT) systems have been shown to outperform their text-only neural machine translation (NMT) counterparts when visual context is available. However, recent studies have also shown that the performance of MMT models is only marginally impacted when the associated image is replaced with an ...
2021.emnlp-main.673
10.18653/v1/2021.emnlp-main.673
null
2109.03415
title_snapshot
2021.emnlp-main.674
Efficient Inference for Multilingual Neural Machine Translation
https://aclanthology.org/2021.emnlp-main.674/
[ "Alexandre Berard", "Dain Lee", "Stephane Clinchant", "Kweonwoo Jung", "Vassilina Nikoulina" ]
Multilingual NMT has become an attractive solution for MT deployment in production. But to match bilingual quality, it comes at the cost of larger and slower models. In this work, we consider several ways to make multilingual NMT faster at inference without degrading its quality. We experiment with several “light decod...
2021.emnlp-main.674
10.18653/v1/2021.emnlp-main.674
null
2109.06679
title_snapshot
2021.emnlp-main.675
Role of Language Relatedness in Multilingual Fine-tuning of Language Models: A Case Study in Indo-Aryan Languages
https://aclanthology.org/2021.emnlp-main.675/
[ "Tejas Dhamecha", "Rudra Murthy", "Samarth Bharadwaj", "Karthik Sankaranarayanan", "Pushpak Bhattacharyya" ]
We explore the impact of leveraging the relatedness of languages that belong to the same family in NLP models using multilingual fine-tuning. We hypothesize and validate that multilingual fine-tuning of pre-trained language models can yield better performance on downstream NLP applications, compared to models fine-tune...
2021.emnlp-main.675
10.18653/v1/2021.emnlp-main.675
null
2109.10534
title_snapshot
2021.emnlp-main.676
Comparing Feature-Engineering and Feature-Learning Approaches for Multilingual Translationese Classification
https://aclanthology.org/2021.emnlp-main.676/
[ "Daria Pylypenko", "Kwabena Amponsah-Kaakyire", "Koel Dutta Chowdhury", "Josef van Genabith", "Cristina España-Bonet" ]
Traditional hand-crafted linguistically-informed features have often been used for distinguishing between translated and original non-translated texts. By contrast, to date, neural architectures without manual feature engineering have been less explored for this task. In this work, we (i) compare the traditional featur...
2021.emnlp-main.676
10.18653/v1/2021.emnlp-main.676
null
2109.07604
title_snapshot
2021.emnlp-main.677
Multi-Sentence Resampling: A Simple Approach to Alleviate Dataset Length Bias and Beam-Search Degradation
https://aclanthology.org/2021.emnlp-main.677/
[ "Ivan Provilkov", "Andrey Malinin" ]
Neural Machine Translation (NMT) is known to suffer from a beam-search problem: after a certain point, increasing beam size causes an overall drop in translation quality. This effect is especially pronounced for long sentences. While much work was done analyzing this phenomenon, primarily for autoregressive NMT models,...
2021.emnlp-main.677
10.18653/v1/2021.emnlp-main.677
null
2109.06253
title_snapshot
2021.emnlp-main.678
Cross-Policy Compliance Detection via Question Answering
https://aclanthology.org/2021.emnlp-main.678/
[ "Marzieh Saeidi", "Majid Yazdani", "Andreas Vlachos" ]
Policy compliance detection is the task of ensuring that a scenario conforms to a policy (e.g. a claim is valid according to government rules or a post in an online platform conforms to community guidelines). This task has been previously instantiated as a form of textual entailment, which results in poor accuracy due ...
2021.emnlp-main.678
10.18653/v1/2021.emnlp-main.678
null
2109.03731
title_snapshot
2021.emnlp-main.679
Meta-LMTC: Meta-Learning for Large-Scale Multi-Label Text Classification
https://aclanthology.org/2021.emnlp-main.679/
[ "Ran Wang", "Xi’ao Su", "Siyu Long", "Xinyu Dai", "Shujian Huang", "Jiajun Chen" ]
Large-scale multi-label text classification (LMTC) tasks often face long-tailed label distributions, where many labels have few or even no training instances. Although current methods can exploit prior knowledge to handle these few/zero-shot labels, they neglect the meta-knowledge contained in the dataset that can guid...
2021.emnlp-main.679
10.18653/v1/2021.emnlp-main.679
null
null
null
2021.emnlp-main.680
Unsupervised Multi-View Post-OCR Error Correction With Language Models
https://aclanthology.org/2021.emnlp-main.680/
[ "Harsh Gupta", "Luciano Del Corro", "Samuel Broscheit", "Johannes Hoffart", "Eliot Brenner" ]
We investigate post-OCR correction in a setting where we have access to different OCR views of the same document. The goal of this study is to understand if a pretrained language model (LM) can be used in an unsupervised way to reconcile the different OCR views such that their combination contains fewer errors than eac...
2021.emnlp-main.680
10.18653/v1/2021.emnlp-main.680
null
null
null
2021.emnlp-main.681
Parallel Refinements for Lexically Constrained Text Generation with BART
https://aclanthology.org/2021.emnlp-main.681/
[ "Xingwei He" ]
Lexically constrained text generation aims to control the generated text by incorporating certain pre-specified keywords into the output. Previous work injects lexical constraints into the output by controlling the decoding process or refining the candidate output iteratively, which tends to generate generic or ungramm...
2021.emnlp-main.681
10.18653/v1/2021.emnlp-main.681
null
2109.12487
title_snapshot
2021.emnlp-main.682
BERT-Beta: A Proactive Probabilistic Approach to Text Moderation
https://aclanthology.org/2021.emnlp-main.682/
[ "Fei Tan", "Yifan Hu", "Kevin Yen", "Changwei Hu" ]
Text moderation for user generated content, which helps to promote healthy interaction among users, has been widely studied and many machine learning models have been proposed. In this work, we explore an alternative perspective by augmenting reactive reviews with proactive forecasting. Specifically, we propose a new c...
2021.emnlp-main.682
10.18653/v1/2021.emnlp-main.682
null
2109.08805
title_snapshot
2021.emnlp-main.683
STaCK: Sentence Ordering with Temporal Commonsense Knowledge
https://aclanthology.org/2021.emnlp-main.683/
[ "Deepanway Ghosal", "Navonil Majumder", "Rada Mihalcea", "Soujanya Poria" ]
Sentence order prediction is the task of finding the correct order of sentences in a randomly ordered document. Correctly ordering the sentences requires an understanding of coherence with respect to the chronological sequence of events described in the text. Document-level contextual understanding and commonsense know...
2021.emnlp-main.683
10.18653/v1/2021.emnlp-main.683
null
2109.02247
title_snapshot
2021.emnlp-main.684
Preventing Author Profiling through Zero-Shot Multilingual Back-Translation
https://aclanthology.org/2021.emnlp-main.684/
[ "David Ifeoluwa Adelani", "Miaoran Zhang", "Xiaoyu Shen", "Ali Davody", "Thomas Kleinbauer", "Dietrich Klakow" ]
Documents as short as a single sentence may inadvertently reveal sensitive information about their authors, including e.g. their gender or ethnicity. Style transfer is an effective way of transforming texts in order to remove any information that enables author profiling. However, for a number of current state-of-the-a...
2021.emnlp-main.684
10.18653/v1/2021.emnlp-main.684
null
2109.09133
title_snapshot
2021.emnlp-main.685
CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generation
https://aclanthology.org/2021.emnlp-main.685/
[ "Yue Wang", "Weishi Wang", "Shafiq Joty", "Steven C.H. Hoi" ]
Pre-trained models for Natural Languages (NL) like BERT and GPT have been recently shown to transfer well to Programming Languages (PL) and largely benefit a broad set of code-related tasks. Despite their success, most current methods either rely on an encoder-only (or decoder-only) pre-training that is suboptimal for ...
2021.emnlp-main.685
10.18653/v1/2021.emnlp-main.685
null
2109.00859
title_snapshot
2021.emnlp-main.686
Detect and Classify – Joint Span Detection and Classification for Health Outcomes
https://aclanthology.org/2021.emnlp-main.686/
[ "Micheal Abaho", "Danushka Bollegala", "Paula Williamson", "Susanna Dodd" ]
A health outcome is a measurement or an observation used to capture and assess the effect of a treatment. Automatic detection of health outcomes from text would undoubtedly speed up access to evidence necessary in healthcare decision making. Prior work on outcome detection has modelled this task as either (a) a sequenc...
2021.emnlp-main.686
10.18653/v1/2021.emnlp-main.686
null
2104.07789
title_snapshot
2021.emnlp-main.687
Multi-Class Grammatical Error Detection for Correction: A Tale of Two Systems
https://aclanthology.org/2021.emnlp-main.687/
[ "Zheng Yuan", "Shiva Taslimipoor", "Christopher Davis", "Christopher Bryant" ]
In this paper, we show how a multi-class grammatical error detection (GED) system can be used to improve grammatical error correction (GEC) for English. Specifically, we first develop a new state-of-the-art binary detection system based on pre-trained ELECTRA, and then extend it to multi-class detection using different...
2021.emnlp-main.687
10.18653/v1/2021.emnlp-main.687
null
null
null
2021.emnlp-main.688
Towards Zero-shot Commonsense Reasoning with Self-supervised Refinement of Language Models
https://aclanthology.org/2021.emnlp-main.688/
[ "Tassilo Klein", "Moin Nabi" ]
Can we get existing language models and refine them for zero-shot commonsense reasoning? This paper presents an initial study exploring the feasibility of zero-shot commonsense reasoning for the Winograd Schema Challenge by formulating the task as self-supervised refinement of a pre-trained language model. In contrast ...
2021.emnlp-main.688
10.18653/v1/2021.emnlp-main.688
null
2109.05105
title_snapshot
2021.emnlp-main.689
To Share or not to Share: Predicting Sets of Sources for Model Transfer Learning
https://aclanthology.org/2021.emnlp-main.689/
[ "Lukas Lange", "Jannik Strötgen", "Heike Adel", "Dietrich Klakow" ]
In low-resource settings, model transfer can help to overcome a lack of labeled data for many tasks and domains. However, predicting useful transfer sources is a challenging problem, as even the most similar sources might lead to unexpected negative transfer results. Thus, ranking methods based on task and text similar...
2021.emnlp-main.689
10.18653/v1/2021.emnlp-main.689
null
2104.08078
title_snapshot
2021.emnlp-main.690
Self-Supervised Detection of Contextual Synonyms in a Multi-Class Setting: Phenotype Annotation Use Case
https://aclanthology.org/2021.emnlp-main.690/
[ "Jingqing Zhang", "Luis Bolanos Trujillo", "Tong Li", "Ashwani Tanwar", "Guilherme Freire", "Xian Yang", "Julia Ive", "Vibhor Gupta", "Yike Guo" ]
Contextualised word embeddings is a powerful tool to detect contextual synonyms. However, most of the current state-of-the-art (SOTA) deep learning concept extraction methods remain supervised and underexploit the potential of the context. In this paper, we propose a self-supervised pre-training approach which is able ...
2021.emnlp-main.690
10.18653/v1/2021.emnlp-main.690
null
2109.01935
title_snapshot
2021.emnlp-main.691
ClauseRec: A Clause Recommendation Framework for AI-aided Contract Authoring
https://aclanthology.org/2021.emnlp-main.691/
[ "Vinay Aggarwal", "Aparna Garimella", "Balaji Vasan Srinivasan", "Anandhavelu N", "Rajiv Jain" ]
Contracts are a common type of legal document that frequent in several day-to-day business workflows. However, there has been very limited NLP research in processing such documents, and even lesser in generating them. These contracts are made up of clauses, and the unique nature of these clauses calls for specific meth...
2021.emnlp-main.691
10.18653/v1/2021.emnlp-main.691
null
2110.15794
title_snapshot
2021.emnlp-main.692
Finnish Dialect Identification: The Effect of Audio and Text
https://aclanthology.org/2021.emnlp-main.692/
[ "Mika Hämäläinen", "Khalid Alnajjar", "Niko Partanen", "Jack Rueter" ]
Finnish is a language with multiple dialects that not only differ from each other in terms of accent (pronunciation) but also in terms of morphological forms and lexical choice. We present the first approach to automatically detect the dialect of a speaker based on a dialect transcript and transcript with audio recordi...
2021.emnlp-main.692
10.18653/v1/2021.emnlp-main.692
null
2111.03800
title_snapshot
2021.emnlp-main.693
English Machine Reading Comprehension Datasets: A Survey
https://aclanthology.org/2021.emnlp-main.693/
[ "Daria Dzendzik", "Jennifer Foster", "Carl Vogel" ]
This paper surveys 60 English Machine Reading Comprehension datasets, with a view to providing a convenient resource for other researchers interested in this problem. We categorize the datasets according to their question and answer form and compare them across various dimensions including size, vocabulary, data source...
2021.emnlp-main.693
10.18653/v1/2021.emnlp-main.693
null
2101.10421
title_snapshot
2021.emnlp-main.694
Expanding End-to-End Question Answering on Differentiable Knowledge Graphs with Intersection
https://aclanthology.org/2021.emnlp-main.694/
[ "Priyanka Sen", "Armin Oliya", "Amir Saffari" ]
End-to-end question answering using a differentiable knowledge graph is a promising technique that requires only weak supervision, produces interpretable results, and is fully differentiable. Previous implementations of this technique (Cohen et al, 2020) have focused on single-entity questions using a relation followin...
2021.emnlp-main.694
10.18653/v1/2021.emnlp-main.694
null
2109.05808
title_snapshot
2021.emnlp-main.695
Structured Context and High-Coverage Grammar for Conversational Question Answering over Knowledge Graphs
https://aclanthology.org/2021.emnlp-main.695/
[ "Pierre Marion", "Pawel Nowak", "Francesco Piccinno" ]
We tackle the problem of weakly-supervised conversational Question Answering over large Knowledge Graphs using a neural semantic parsing approach. We introduce a new Logical Form (LF) grammar that can model a wide range of queries on the graph while remaining sufficiently simple to generate supervision data efficiently...
2021.emnlp-main.695
10.18653/v1/2021.emnlp-main.695
null
2109.00269
title_snapshot
2021.emnlp-main.696
Improving Question Answering Model Robustness with Synthetic Adversarial Data Generation
https://aclanthology.org/2021.emnlp-main.696/
[ "Max Bartolo", "Tristan Thrush", "Robin Jia", "Sebastian Riedel", "Pontus Stenetorp", "Douwe Kiela" ]
Despite recent progress, state-of-the-art question answering models remain vulnerable to a variety of adversarial attacks. While dynamic adversarial data collection, in which a human annotator tries to write examples that fool a model-in-the-loop, can improve model robustness, this process is expensive which limits the...
2021.emnlp-main.696
10.18653/v1/2021.emnlp-main.696
null
2104.08678
title_snapshot
2021.emnlp-main.697
BeliefBank: Adding Memory to a Pre-Trained Language Model for a Systematic Notion of Belief
https://aclanthology.org/2021.emnlp-main.697/
[ "Nora Kassner", "Oyvind Tafjord", "Hinrich Schütze", "Peter Clark" ]
Although pretrained language models (PTLMs) contain significant amounts of world knowledge, they can still produce inconsistent answers to questions when probed, even after specialized training. As a result, it can be hard to identify what the model actually “believes” about the world, making it susceptible to inconsis...
2021.emnlp-main.697
10.18653/v1/2021.emnlp-main.697
null
2109.14723
title_snapshot
2021.emnlp-main.698
MLEC-QA: A Chinese Multi-Choice Biomedical Question Answering Dataset
https://aclanthology.org/2021.emnlp-main.698/
[ "Jing Li", "Shangping Zhong", "Kaizhi Chen" ]
Question Answering (QA) has been successfully applied in scenarios of human-computer interaction such as chatbots and search engines. However, for the specific biomedical domain, QA systems are still immature due to expert-annotated datasets being limited by category and scale. In this paper, we present MLEC-QA, the la...
2021.emnlp-main.698
10.18653/v1/2021.emnlp-main.698
null
null
null
2021.emnlp-main.699
IndoNLG: Benchmark and Resources for Evaluating Indonesian Natural Language Generation
https://aclanthology.org/2021.emnlp-main.699/
[ "Samuel Cahyawijaya", "Genta Indra Winata", "Bryan Wilie", "Karissa Vincentio", "Xiaohong Li", "Adhiguna Kuncoro", "Sebastian Ruder", "Zhi Yuan Lim", "Syafri Bahar", "Masayu Khodra", "Ayu Purwarianti", "Pascale Fung" ]
Natural language generation (NLG) benchmarks provide an important avenue to measure progress and develop better NLG systems. Unfortunately, the lack of publicly available NLG benchmarks for low-resource languages poses a challenging barrier for building NLG systems that work well for languages with limited amounts of d...
2021.emnlp-main.699
10.18653/v1/2021.emnlp-main.699
null
2104.08200
title_snapshot
2021.emnlp-main.700
Is Multi-Hop Reasoning Really Explainable? Towards Benchmarking Reasoning Interpretability
https://aclanthology.org/2021.emnlp-main.700/
[ "Xin Lv", "Yixin Cao", "Lei Hou", "Juanzi Li", "Zhiyuan Liu", "Yichi Zhang", "Zelin Dai" ]
Multi-hop reasoning has been widely studied in recent years to obtain more interpretable link prediction. However, we find in experiments that many paths given by these models are actually unreasonable, while little work has been done on interpretability evaluation for them. In this paper, we propose a unified framewor...
2021.emnlp-main.700
10.18653/v1/2021.emnlp-main.700
null
2104.06751
title_snapshot