OneStopEnglish corpus: A new corpus for automatic readability
            assessment and text simplification

Vajjala, Sowmya; Lučić, Ivana

doi:10.18653/v1/w18-0535

Cited by 96 publications

(97 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…It was already noted by (Vajjala and Meurers, 2012) that the actual readability level of each test was difficult to predict accurately based on the sole FKGL. OSE complexity levels are also in agreement with the Flesch-Kincaid index, and in agreement with the numbers reported in (Vajjala and Lucic, 2018); again we see a large overlap between levels for this index. Overall, OSE texts are somewhat more complex than Weebit's, with OSE level 1 comparable in difficulty to Weebit level 4.…”

Section: Sourcesupporting

confidence: 90%

Measuring text readability with machine comprehension: a pilot study

Benzahra¹,

Yvon²

2019

Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications

View full text Add to dashboard Cite

This article studies the relationship between text readability levels and automatic machine understanding systems. Our hypothesis is that the simpler a text is, the better it should be understood by a machine. We thus expect a strong correlation between readability levels on the one hand, and performance of automatic reading systems on the other hand. We test this hypothesis with several understanding systems based on language models of varying strengths, measuring this correlation on two corpora of journalistic texts. Our results suggest that this correlation is quite small and that existing comprehension systems are far to reproduce the gradual improvement of their performance on texts of decreasing complexity.

show abstract

Section: Sourcesupporting

confidence: 90%

Measuring text readability with machine comprehension: a pilot study

Benzahra¹,

Yvon²

2019

Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications

View full text Add to dashboard Cite

show abstract

“…However, we know it is possible to automatically distinguish between these levels in this corpus using machine learning models (Ambati et al, 2016;Vajjala and Lucic, 2018). Whether the variation between texts of any specific linguistic property (e.g., lexical richness, syntactic complexity, coherence) can be correlated with the differences in comprehension scores instead of "reading level" assigned by the teachers should be explored as a part of future work.…”

Section: Discussionmentioning

confidence: 99%

“…Texts: We randomly selected 15 texts from the OneStopEnglish corpus (Vajjala and Lucic, 2018), consisting of manually simplified news articles from The Guardian, by English teachers, to suit beginner, intermediate, and advanced readers of English as Second Language (ESL). This corpus was also used in past user studies related to readability assessment (Crossley et al, 2014; Participants: 112 non-native English speaking participants were recruited for this study from among the student population of an American university by means of an internal email advertisement.…”

Section: Methods and Experiments Proceduresmentioning

confidence: 99%

On Understanding the Relation between Expert Annotations of Text Readability and Target Reader Comprehension

Vajjala

Lučić²

2019

Proceedings of the Fourteenth Workshop on Innovative Use of NLP for Building Educational Applications

Self Cite

View full text Add to dashboard Cite

Automatic readability assessment aims to ensure that readers read texts that they can comprehend. However, computational models are typically trained on texts created from the perspective of the text writer, not the target reader. There is little experimental research on the relationship between expert annotations of readability, reader's language proficiency, and different levels of reading comprehension. To address this gap, we conducted a user study in which over a 100 participants read texts of different reading levels and answered questions created to test three forms of comprehension. Our results indicate that more than readability annotation or reader proficiency, it is the type of comprehension question asked that shows differences between reader responses-inferential questions were difficult for users of all levels of proficiency across reading levels. The data collected from this study is released with this paper 1 , which will, for the first time, provide a collection of 45 reader bench marked texts to evaluate readability assessment systems developed for adult learners of English. It can also potentially be useful for the development of question generation approaches in intelligent tutoring systems research.

show abstract

“…This advantage of this method lies in its high reliability, but inviting the experts is carefully considered. This was also the method of Slyh & Hansen [15], Todirascu et al [16], Xia et al [17], Vajjala & Lučić [18], Nguyen & Henkin [8,9], among others.…”

Section: Criteria For Building the Corpusmentioning

confidence: 99%

Building a Corpus for Vietnamese Text Readability Assessment in The Literature Domain

Luong¹,

Nguyen²,

Đinh³

2020

ujer

View full text Add to dashboard Cite

Text readability is a measure of how easy or difficult it is to read a text. This readability factor plays a crucial role in the processes of drafting and comprehending the texts, affecting the choice of proper texts for reading. Studies on the readability of text have started since the late nineteenth century and there have been many practical applications. However, these studies are mainly performed in English and other popular languages. In Vietnamese, the study of the text readability is still relatively untapped and has only received attention in recent years in the process of improving the curriculum and teaching methods. Recent studies on the readability of text in Vietnamese language are still limited, the main reason was largely due to the lack of text resources, which are corpora graded accordingly to difficulty levels. Therefore, in this study, we focused on building a corpus for assessing the readability of Vietnamese texts in the literature domain through the process of collecting, processing and evaluating documents. The result is that we have built up a corpus of 1,825 Vietnamese texts, divided into four levels of difficulty (Very easy, Easy, Medium and Difficult). Experiments with the existing Vietnamese readability assessment methods show that the built corpus is reliable and usable for further research on the text readability.

show abstract

OneStopEnglish corpus: A new corpus for automatic readability assessment and text simplification

Cited by 96 publications

References 26 publications

Measuring text readability with machine comprehension: a pilot study

Measuring text readability with machine comprehension: a pilot study

On Understanding the Relation between Expert Annotations of Text Readability and Target Reader Comprehension

Building a Corpus for Vietnamese Text Readability Assessment in The Literature Domain

Contact Info

Product

Resources

About