Construction of a Russian Paraphrase Corpus: Unsupervised Paraphrase Extraction

Pronoza, Ekaterina V.; Yagunova, Elena; Pronoza, Anton

doi:10.1007/978-3-319-41718-9_8

Cited by 22 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For the above mentioned tasks we are going to use the following corpora: Paraphraser.ru (Pronoza et al, 2016) for the Russian language paraphrase identification task, Microsoft Research Paraphrase Corpus (Dolan et al, 2004) for the English language paraphrase identification task, Turkish Paraphrase Corpus (Demir et al, 2012) for the Turkish language paraphrase identification task; Russian Twitter Sentiment Corpus (Rubtsova, 2014) for the Russian language sentiment analysis task, Stanford Sentiment Treebank (Socher et al, 2013) for the English language sentiment analysis task; and Stanford Natural Language Inference (Bowman et al, 2015) for the English language natural language inference task.…”

Section: Methodsmentioning

confidence: 99%

Robust to Noise Models in Natural Language Processing Tasks

Malykh

2019

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop

View full text Add to dashboard Cite

There are a lot of noisy texts surrounding a person in modern life. A traditional approach is to use spelling correction, yet the existing solutions are far from perfect. We propose a robust to noise word embeddings model which outperforms existing commonly used models like fasttext and word2vec in different tasks. In addition, we investigate the noise robustness of current models in different natural language processing tasks. We propose extensions for modern models in three downstream tasks, i.e. text classification, named entity recognition and aspect extraction, these extensions show improvement in noise robustness over existing solutions.

show abstract

Section: Methodsmentioning

confidence: 99%

Robust to Noise Models in Natural Language Processing Tasks

Malykh

2019

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop

View full text Add to dashboard Cite

show abstract

“…Paraphraser [17] is a dataset for the paraphrasing task: it consists of sentence pairs, each of which is labeled as paraphrase, not paraphrase or maybe paraphrase. This task is close to DaNetQA as the model is required to detect linkage between sentences.…”

Section: Task Transferringmentioning

confidence: 99%

DaNetQA: A Yes/No Question Answering Dataset for the Russian Language

Глушкова

Machnev

Fenogenova

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

DaNetQA, a new question-answering corpus, follows BoolQ [2] design: it comprises natural yes/no questions. Each question is paired with a paragraph from Wikipedia and an answer, derived from the paragraph. The task is to take both the question and a paragraph as input and come up with a yes/no answer, i.e. to produce a binary output. In this paper, we present a reproducible approach to DaNetQA creation and investigate transfer learning methods for task and language transferring. For task transferring we leverage three similar sentence modelling tasks: 1) a corpus of paraphrases, Paraphraser, 2) an NLI task, for which we use the Russian part of XNLI, 3) another question answering task, SberQUAD. For language transferring we use English to Russian translation together with multilingual language fine-tuning.

show abstract

“…• English -Microsoft Research Paraphrase Corpus (Dolan et al, 2004) consists of 5,800 sentence pairs extracted from news sources on the web and manually labelled for presence/absence of semantic equivalence. • Russian -Russian Paraphrase Corpus (Pronoza et al, 2016) consists of news headings from different news agencies. It contains around 6,000 pairs of phrases labelled in terms of ternary scale: "-1" -not paraphrase, "0" -weak paraphrase, and "1"strong paraphrase.…”

Section: Paraphrase Detectionmentioning

confidence: 99%

Robust Word Vectors: Context-Informed Embeddings for Noisy Texts

Malykh¹,

Logacheva²,

Khakhulin³

2018

Proceedings of the 2018 EMNLP Workshop W-Nut: The 4th Workshop on Noisy User-Generated Text

View full text Add to dashboard Cite

We suggest a new language-independent architecture of robust word vectors (RoVe). It is designed to alleviate the issue of typos, which are common in almost any user-generated content, and hinder automatic text processing. Our model is morphologically motivated, which allows it to deal with unseen word forms in morphologically rich languages. We present the results on a number of Natural Language Processing (NLP) tasks and languages for the variety of related architectures and show that proposed architecture is typo-proof.

show abstract

Construction of a Russian Paraphrase Corpus: Unsupervised Paraphrase Extraction

Cited by 22 publications

References 14 publications

Robust to Noise Models in Natural Language Processing Tasks

Robust to Noise Models in Natural Language Processing Tasks

DaNetQA: A Yes/No Question Answering Dataset for the Russian Language

Robust Word Vectors: Context-Informed Embeddings for Noisy Texts

Contact Info

Product

Resources

About