LREC 2020

230 papers
The 2020 edition of the Language Resources and Evaluation Conference (LREC) featured 241 papers focusing on expanding natural language processing capabilities to underserved languages and specialized domains. A primary theme was addressing linguistic inequality, with research dedicated to neural machine translation for low-resource languages and scaling models to the global long tail. Other prominent topics included subword representations for robust language understanding, language-specific pretraining to capture local linguistic nuances, and the development of neural representations for morphologically rich languages. Contributions included the release of specialized datasets and toolkits, such as MultiWOZ 2.1 for multi-domain dialogue, the ParaPat parallel corpus of patent abstracts, and Seq2SeqPy for sequence-to-sequence modeling. Research also explored orthographic normalization for non-standardized dialects and the processing of physically degraded historical documents. Prolific contributors to this year’s proceedings included Benoît Sagot, Nizar Habash, and Rada Mihalcea. The year's work reflected a trend toward cross-domain sentiment analysis and the creation of domain-specific resources for fields like archaeology and pedagogy.

Top Research Directions

Research directions most represented by LREC 2020 papers.

The Performance Advantage of Localized Language Pretraining 9
Learning from Noisy Distantly Supervised Labels 7
Orthographic Normalization for Dialectal and Informal Text 6
Character N-Gram Embeddings for Robust Word Representations 6
Orthographic Inconsistency in Low Resource Language Processing 6
Resolving Domain Specific Sentiment Shift 5
Automating Lexical Semantic Resource Construction 5
Overcoming Vocabulary Sparsity in Morphologically Rich Languages 5
Scaling Training Data via Distant Supervision 4
Pointer Generator Mechanisms for Grounded Text Generation 4
Resolving Zero Anaphora in Pro Drop Languages 4
The Mechanics of Subword Tokenization 4
Resolving the Semantic Conflation Bottleneck in Word Embeddings 4
Modeling Semantic Inversion in Sarcasm Detection 4
Robustness to Formal and Informal Domain Shift 4

Top Authors

AuthorPapers
Nizar Habash5
Arya D. McCarthy4
Benoît Sagot4
Franck Dernoncourt4
Rada Mihalcea4
Aaron Mueller3
Anders Søgaard3
Jörg Tiedemann3
Lilja Øvrelid3
Oier Lopez de Lacalle3
Peter Jansen3
Salam Khalifa3
Winston Wu3
Albert Gatt2
Boi Faltings2
Clare R. Voss2
Dan Roth2
Diego Antognini2
Diptesh Kanojia2
Eneko Agirre2

Impactful Papers

Papers from LREC 2020 that shaped research clusters.

Foundational

Influential

Papers

Title
Temporal Histories of Epidemic Events (THEE): A Case Study in Temporal Annotation for Public Health
ThaiLMCut: Unsupervised Pretraining for Thai Word Segmentation
The Connection between the Text and Images of News Articles: New Insights for Multimedia Analysis
The Design and Construction of a Chinese Sarcasm Dataset
The European Language Technology Landscape in 2020: Language-Centric and Human-Centric AI for Cross-Cultural Communication in Multilingual Europe
The FISKMÖ Project: Resources and Tools for Finnish-Swedish Machine Translation and Cross-Linguistic Research
The JDDC Corpus: A Large-Scale Multi-Turn Chinese Dialogue Dataset for E-commerce Customer Service
The Johns Hopkins University Bible Corpus: 1600+ Tongues for Typological Exploration
The Margarita Dialogue Corpus: A Data Set for Time-Offset Interactions and Unstructured Dialogue Systems
The Medical Scribe: Corpus Development and Model Performance Analyses
The POTUS Corpus, a Database of Weekly Addresses for the Study of Stance in Politics and Virtual Agents
The rJokes Dataset: a Large Scale Humor Collection
The Universal Decompositional Semantics Dataset and Decomp Toolkit
Towards the Necessity for Debiasing Natural Language Inference Datasets
TutorialVQA: Question Answering Dataset for Tutorial Videos
Understanding Spatial Relations through Multiple Modalities
UniMorph 3.0: Universal Morphology
Universal Dependencies v2: An Evergrowing Multilingual Treebank Collection
Unsupervised Domain Adaptation of Language Models for Reading Comprehension
Using Deep Neural Networks with Intra- and Inter-Sentence Context to Classify Suicidal Behaviour
Using Distributional Thesaurus Embedding for Co-hyponymy Detection
Visual Grounding Annotation of Recipe Flow Graph
Which Evaluations Uncover Sense Representations that Actually Make Sense?
Wiki-40B: Multilingual Language Model Dataset
WikiBank: Using Wikidata to Improve Multilingual Frame-Semantic Parsing
WIKIR: A Python Toolkit for Building a Large-scale Wikipedia-based English Information Retrieval Dataset
WN-Salience: A Corpus of News Articles with Entity Salience Annotations
word2word: A Collection of Bilingual Lexicons for 3, 564 Language Pairs
Word Sense Disambiguation for 158 Languages using Word Embeddings Only
WorldTree V2: A Corpus of Science-Domain Structured Explanations and Inference Patterns supporting Multi-Hop Inference

« prev Page 5/5