Dataset Viewer
Full Screen Viewer
Full Screen
The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
Dataset Card for MultiSim Benchmark
Dataset Summary
The MultiSim benchmark is a growing collection of text simplification datasets targeted at sentence simplification in several languages. Currently, the benchmark spans 12 languages.
Supported Tasks
- Sentence Simplification
Usage
from datasets import load_dataset
dataset = load_dataset("MichaelR207/MultiSim")
Citation
If you use this benchmark, please cite our paper:
@inproceedings{ryan-etal-2023-revisiting,
title = "Revisiting non-{E}nglish Text Simplification: A Unified Multilingual Benchmark",
author = "Ryan, Michael and
Naous, Tarek and
Xu, Wei",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.269",
pages = "4898--4927",
abstract = "Recent advancements in high-quality, large-scale English resources have pushed the frontier of English Automatic Text Simplification (ATS) research. However, less work has been done on multilingual text simplification due to the lack of a diverse evaluation benchmark that covers complex-simple sentence pairs in many languages. This paper introduces the MultiSim benchmark, a collection of 27 resources in 12 distinct languages containing over 1.7 million complex-simple sentence pairs. This benchmark will encourage research in developing more effective multilingual text simplification models and evaluation metrics. Our experiments using MultiSim with pre-trained multilingual language models reveal exciting performance improvements from multilingual training in non-English settings. We observe strong performance from Russian in zero-shot cross-lingual transfer to low-resource languages. We further show that few-shot prompting with BLOOM-176b achieves comparable quality to reference simplifications outperforming fine-tuned models in most languages. We validate these findings through human evaluation.",
}
Contact
Michael Ryan: Scholar | Twitter | Github | LinkedIn | Research Gate | Personal Website | michaeljryan@stanford.edu
Languages
- English
- French
- Russian
- Japanese
- Italian
- Danish (on request)
- Spanish (on request)
- German
- Brazilian Portuguese
- Slovene
- Urdu (on request)
- Basque (on request)
Dataset Structure
Data Instances
MultiSim is a collection of 27 existing datasets:
- AdminIT
- ASSET
- CBST
- CLEAR
- DSim
- Easy Japanese
- Easy Japanese Extended
- GEOLino
- German News
- Newsela EN/ES
- PaCCSS-IT
- PorSimples
- RSSE
- RuAdapt Encyclopedia
- RuAdapt Fairytales
- RuAdapt Literature
- RuWikiLarge
- SIMPITIKI
- Simple German
- Simplext
- SimplifyUR
- SloTS
- Teacher
- Terence
- TextComplexityDE
- WikiAuto
- WikiLargeFR
Data Fields
In the train set, you will only find original and simple sentences. In the validation and test sets you may find simple1, simple2, ... simpleN because a given sentence can have multiple reference simplifications (useful in SARI and BLEU calculations)
Data Splits
The dataset is split into a train, validation, and test set.
Dataset Creation
Curation Rationale
I hope that collecting all of these independently useful resources for text simplification together into one benchmark will encourage multilingual work on text simplification!
Source Data
Initial Data Collection and Normalization
Data is compiled from the 27 existing datasets that comprise the MultiSim Benchmark. For details on each of the resources please see Appendix A in the paper.
Who are the source language producers?
Each dataset has different sources. At a high level the sources are: Automatically Collected (ex. Wikipedia, Web data), Manually Collected (ex. annotators asked to simplify sentences), Target Audience Resources (ex. Newsela News Articles), or Translated (ex. Machine translations of existing datasets).
These sources can be seen in Table 1 pictured above (Section: Dataset Structure/Data Instances) and further discussed in section 3 of the paper. Appendix A of the paper has details on specific resources.
Annotations
Annotation process
Annotators writing simplifications (only for some datasets) typically follow an annotation guideline. Some example guidelines come from here, here, and here.
Who are the annotators?
See Table 1 (Section: Dataset Structure/Data Instances) for specific annotators per dataset. At a high level the annotators are: writers, translators, teachers, linguists, journalists, crowdworkers, experts, news agencies, medical students, students, writers, and researchers.
Personal and Sensitive Information
No dataset should contain personal or sensitive information. These were previously collected resources primarily collected from news sources, wikipedia, science communications, etc. and were not identified to have personally identifiable information.
Considerations for Using the Data
Social Impact of Dataset
We hope this dataset will make a greatly positive social impact as text simplification is a task that serves children, second language learners, and people with reading/cognitive disabilities. By publicly releasing a dataset in 12 languages we hope to serve these global communities. One negative and unintended use case for this data would be reversing the labels to make a "text complification" model. We beleive the benefits of releasing this data outweigh the harms and hope that people use the dataset as intended.
Discussion of Biases
There may be biases of the annotators involved in writing the simplifications towards how they believe a simpler sentence should be written. Additionally annotators and editors have the choice of what information does not make the cut in the simpler sentence introducing information importance bias.
Other Known Limitations
Some of the included resources were automatically collected or machine translated. As such not every sentence is perfectly aligned. Users are recommended to use such individual resources with caution.
Additional Information
Dataset Curators
Michael Ryan: Scholar | Twitter | Github | LinkedIn | Research Gate | Personal Website | michaeljryan@stanford.edu
Licensing Information
MIT License
Citation Information
Please cite the individual datasets that you use within the MultiSim benchmark as appropriate. Proper bibtex attributions for each of the datasets are included below.
AdminIT
@inproceedings{miliani-etal-2022-neural,
title = "Neural Readability Pairwise Ranking for Sentences in {I}talian Administrative Language",
author = "Miliani, Martina and
Auriemma, Serena and
Alva-Manchego, Fernando and
Lenci, Alessandro",
booktitle = "Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing",
month = nov,
year = "2022",
address = "Online only",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.aacl-main.63",
pages = "849--866",
abstract = "Automatic Readability Assessment aims at assigning a complexity level to a given text, which could help improve the accessibility to information in specific domains, such as the administrative one. In this paper, we investigate the behavior of a Neural Pairwise Ranking Model (NPRM) for sentence-level readability assessment of Italian administrative texts. To deal with data scarcity, we experiment with cross-lingual, cross- and in-domain approaches, and test our models on Admin-It, a new parallel corpus in the Italian administrative language, containing sentences simplified using three different rewriting strategies. We show that NPRMs are effective in zero-shot scenarios ({\textasciitilde}0.78 ranking accuracy), especially with ranking pairs containing simplifications produced by overall rewriting at the sentence-level, and that the best results are obtained by adding in-domain data (achieving perfect performance for such sentence pairs). Finally, we investigate where NPRMs failed, showing that the characteristics of the training data, rather than its size, have a bigger effect on a model{'}s performance.",
}
ASSET
@inproceedings{alva-manchego-etal-2020-asset,
title = "{ASSET}: {A} Dataset for Tuning and Evaluation of Sentence Simplification Models with Multiple Rewriting Transformations",
author = "Alva-Manchego, Fernando and
Martin, Louis and
Bordes, Antoine and
Scarton, Carolina and
Sagot, Beno{\^\i}t and
Specia, Lucia",
booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
month = jul,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2020.acl-main.424",
pages = "4668--4679",
}
CBST
@article{10.1007/s10579-017-9407-6,
title={{The corpus of Basque simplified texts (CBST)}},
author={Gonzalez-Dios, Itziar and Aranzabe, Mar{\'\i}a Jes{\'u}s and D{\'\i}az de Ilarraza, Arantza},
journal={Language Resources and Evaluation},
volume={52},
number={1},
pages={217--247},
year={2018},
publisher={Springer}
}
CLEAR
@inproceedings{grabar-cardon-2018-clear,
title = "{CLEAR} {--} Simple Corpus for Medical {F}rench",
author = "Grabar, Natalia and
Cardon, R{\'e}mi",
booktitle = "Proceedings of the 1st Workshop on Automatic Text Adaptation ({ATA})",
month = nov,
year = "2018",
address = "Tilburg, the Netherlands",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/W18-7002",
doi = "10.18653/v1/W18-7002",
pages = "3--9",
}
DSim
@inproceedings{klerke-sogaard-2012-dsim,
title = "{DS}im, a {D}anish Parallel Corpus for Text Simplification",
author = "Klerke, Sigrid and
S{\o}gaard, Anders",
booktitle = "Proceedings of the Eighth International Conference on Language Resources and Evaluation ({LREC}'12)",
month = may,
year = "2012",
address = "Istanbul, Turkey",
publisher = "European Language Resources Association (ELRA)",
url = "http://www.lrec-conf.org/proceedings/lrec2012/pdf/270_Paper.pdf",
pages = "4015--4018",
abstract = "We present DSim, a new sentence aligned Danish monolingual parallel corpus extracted from 3701 pairs of news telegrams and corresponding professionally simplified short news articles. The corpus is intended for building automatic text simplification for adult readers. We compare DSim to different examples of monolingual parallel corpora, and we argue that this corpus is a promising basis for future development of automatic data-driven text simplification systems in Danish. The corpus contains both the collection of paired articles and a sentence aligned bitext, and we show that sentence alignment using simple tf*idf weighted cosine similarity scoring is on line with state―of―the―art when evaluated against a hand-aligned sample. The alignment results are compared to state of the art for English sentence alignment. We finally compare the source and simplified sides of the corpus in terms of lexical and syntactic characteristics and readability, and find that the one―to―many sentence aligned corpus is representative of the sentence simplifications observed in the unaligned collection of article pairs.",
}
Easy Japanese
@inproceedings{maruyama-yamamoto-2018-simplified,
title = "Simplified Corpus with Core Vocabulary",
author = "Maruyama, Takumi and
Yamamoto, Kazuhide",
booktitle = "Proceedings of the Eleventh International Conference on Language Resources and Evaluation ({LREC} 2018)",
month = may,
year = "2018",
address = "Miyazaki, Japan",
publisher = "European Language Resources Association (ELRA)",
url = "https://aclanthology.org/L18-1185",
}
Easy Japanese Extended
@inproceedings{katsuta-yamamoto-2018-crowdsourced,
title = "Crowdsourced Corpus of Sentence Simplification with Core Vocabulary",
author = "Katsuta, Akihiro and
Yamamoto, Kazuhide",
booktitle = "Proceedings of the Eleventh International Conference on Language Resources and Evaluation ({LREC} 2018)",
month = may,
year = "2018",
address = "Miyazaki, Japan",
publisher = "European Language Resources Association (ELRA)",
url = "https://aclanthology.org/L18-1072",
}
GEOLino
@inproceedings{mallinson2020,
title={Zero-Shot Crosslingual Sentence Simplification},
author={Mallinson, Jonathan and Sennrich, Rico and Lapata, Mirella},
year={2020},
booktitle={2020 Conference on Empirical Methods in Natural Language Processing (EMNLP 2020)}
}
German News
@inproceedings{sauberli-etal-2020-benchmarking,
title = "Benchmarking Data-driven Automatic Text Simplification for {G}erman",
author = {S{\"a}uberli, Andreas and
Ebling, Sarah and
Volk, Martin},
booktitle = "Proceedings of the 1st Workshop on Tools and Resources to Empower People with REAding DIfficulties (READI)",
month = may,
year = "2020",
address = "Marseille, France",
publisher = "European Language Resources Association",
url = "https://aclanthology.org/2020.readi-1.7",
pages = "41--48",
abstract = "Automatic text simplification is an active research area, and there are first systems for English, Spanish, Portuguese, and Italian. For German, no data-driven approach exists to this date, due to a lack of training data. In this paper, we present a parallel corpus of news items in German with corresponding simplifications on two complexity levels. The simplifications have been produced according to a well-documented set of guidelines. We then report on experiments in automatically simplifying the German news items using state-of-the-art neural machine translation techniques. We demonstrate that despite our small parallel corpus, our neural models were able to learn essential features of simplified language, such as lexical substitutions, deletion of less relevant words and phrases, and sentence shortening.",
language = "English",
ISBN = "979-10-95546-45-0",
}
Newsela EN/ES
@article{xu-etal-2015-problems,
title = "Problems in Current Text Simplification Research: New Data Can Help",
author = "Xu, Wei and
Callison-Burch, Chris and
Napoles, Courtney",
journal = "Transactions of the Association for Computational Linguistics",
volume = "3",
year = "2015",
address = "Cambridge, MA",
publisher = "MIT Press",
url = "https://aclanthology.org/Q15-1021",
doi = "10.1162/tacl_a_00139",
pages = "283--297",
abstract = "Simple Wikipedia has dominated simplification research in the past 5 years. In this opinion paper, we argue that focusing on Wikipedia limits simplification research. We back up our arguments with corpus analysis and by highlighting statements that other researchers have made in the simplification literature. We introduce a new simplification dataset that is a significant improvement over Simple Wikipedia, and present a novel quantitative-comparative approach to study the quality of simplification data resources.",
}
PaCCSS-IT
@inproceedings{brunato-etal-2016-paccss,
title = "{P}a{CCSS}-{IT}: A Parallel Corpus of Complex-Simple Sentences for Automatic Text Simplification",
author = "Brunato, Dominique and
Cimino, Andrea and
Dell{'}Orletta, Felice and
Venturi, Giulia",
booktitle = "Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2016",
address = "Austin, Texas",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/D16-1034",
doi = "10.18653/v1/D16-1034",
pages = "351--361",
}
PorSimples
@inproceedings{aluisio-gasperin-2010-fostering,
title = "Fostering Digital Inclusion and Accessibility: The {P}or{S}imples project for Simplification of {P}ortuguese Texts",
author = "Alu{\'\i}sio, Sandra and
Gasperin, Caroline",
booktitle = "Proceedings of the {NAACL} {HLT} 2010 Young Investigators Workshop on Computational Approaches to Languages of the {A}mericas",
month = jun,
year = "2010",
address = "Los Angeles, California",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/W10-1607",
pages = "46--53",
}
@inproceedings{10.1007/978-3-642-16952-6_31,
author="Scarton, Carolina and Gasperin, Caroline and Aluisio, Sandra",
editor="Kuri-Morales, Angel and Simari, Guillermo R.",
title="Revisiting the Readability Assessment of Texts in Portuguese",
booktitle="Advances in Artificial Intelligence -- IBERAMIA 2010",
year="2010",
publisher="Springer Berlin Heidelberg",
address="Berlin, Heidelberg",
pages="306--315",
isbn="978-3-642-16952-6"
}
RSSE
@inproceedings{sakhovskiy2021rusimplesenteval,
title={{RuSimpleSentEval-2021 shared task:} evaluating sentence simplification for Russian},
author={Sakhovskiy, Andrey and Izhevskaya, Alexandra and Pestova, Alena and Tutubalina, Elena and Malykh, Valentin and Smurov, Ivana and Artemova, Ekaterina},
booktitle={Proceedings of the International Conference “Dialogue},
pages={607--617},
year={2021}
}
RuAdapt
@inproceedings{Dmitrieva2021Quantitative,
title={A quantitative study of simplification strategies in adapted texts for L2 learners of Russian},
author={Dmitrieva, Anna and Laposhina, Antonina and Lebedeva, Maria},
booktitle={Proceedings of the International Conference “Dialogue},
pages={191--203},
year={2021}
}
@inproceedings{dmitrieva-tiedemann-2021-creating,
title = "Creating an Aligned {R}ussian Text Simplification Dataset from Language Learner Data",
author = {Dmitrieva, Anna and
Tiedemann, J{\"o}rg},
booktitle = "Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing",
month = apr,
year = "2021",
address = "Kiyv, Ukraine",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.bsnlp-1.8",
pages = "73--79",
abstract = "Parallel language corpora where regular texts are aligned with their simplified versions can be used in both natural language processing and theoretical linguistic studies. They are essential for the task of automatic text simplification, but can also provide valuable insights into the characteristics that make texts more accessible and reveal strategies that human experts use to simplify texts. Today, there exist a few parallel datasets for English and Simple English, but many other languages lack such data. In this paper we describe our work on creating an aligned Russian-Simple Russian dataset composed of Russian literature texts adapted for learners of Russian as a foreign language. This will be the first parallel dataset in this domain, and one of the first Simple Russian datasets in general.",
}
RuWikiLarge
@inproceedings{sakhovskiy2021rusimplesenteval,
title={{RuSimpleSentEval-2021 shared task:} evaluating sentence simplification for Russian},
author={Sakhovskiy, Andrey and Izhevskaya, Alexandra and Pestova, Alena and Tutubalina, Elena and Malykh, Valentin and Smurov, Ivana and Artemova, Ekaterina},
booktitle={Proceedings of the International Conference “Dialogue},
pages={607--617},
year={2021}
}
SIMPITIKI
@article{tonelli2016simpitiki,
title={SIMPITIKI: a Simplification corpus for Italian},
author={Tonelli, Sara and Aprosio, Alessio Palmero and Saltori, Francesca},
journal={Proceedings of CLiC-it},
year={2016}
}
Simple German
@inproceedings{battisti-etal-2020-corpus,
title = "A Corpus for Automatic Readability Assessment and Text Simplification of {G}erman",
author = {Battisti, Alessia and
Pf{\"u}tze, Dominik and
S{\"a}uberli, Andreas and
Kostrzewa, Marek and
Ebling, Sarah},
booktitle = "Proceedings of the Twelfth Language Resources and Evaluation Conference",
month = may,
year = "2020",
address = "Marseille, France",
publisher = "European Language Resources Association",
url = "https://aclanthology.org/2020.lrec-1.404",
pages = "3302--3311",
abstract = "In this paper, we present a corpus for use in automatic readability assessment and automatic text simplification for German, the first of its kind for this language. The corpus is compiled from web sources and consists of parallel as well as monolingual-only (simplified German) data amounting to approximately 6,200 documents (nearly 211,000 sentences). As a unique feature, the corpus contains information on text structure (e.g., paragraphs, lines), typography (e.g., font type, font style), and images (content, position, and dimensions). While the importance of considering such information in machine learning tasks involving simplified language, such as readability assessment, has repeatedly been stressed in the literature, we provide empirical evidence for its benefit. We also demonstrate the added value of leveraging monolingual-only data for automatic text simplification via machine translation through applying back-translation, a data augmentation technique.",
language = "English",
ISBN = "979-10-95546-34-4",
}
Simplext
@article{10.1145/2738046,
author = {Saggion, Horacio and \v{S}tajner, Sanja and Bott, Stefan and Mille, Simon and Rello, Luz and Drndarevic, Biljana},
title = {Making It Simplext: Implementation and Evaluation of a Text Simplification System for Spanish},
year = {2015},
issue_date = {June 2015}, publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
volume = {6},
number = {4},
issn = {1936-7228},
url = {https://doi.org/10.1145/2738046},
doi = {10.1145/2738046},
journal = {ACM Trans. Access. Comput.},
month = {may},
articleno = {14},
numpages = {36},
keywords = {Spanish, text simplification corpus, human evaluation, readability measures}
}
SimplifyUR
@inproceedings{qasmi-etal-2020-simplifyur,
title = "{S}implify{UR}: Unsupervised Lexical Text Simplification for {U}rdu",
author = "Qasmi, Namoos Hayat and
Zia, Haris Bin and
Athar, Awais and
Raza, Agha Ali",
booktitle = "Proceedings of the Twelfth Language Resources and Evaluation Conference",
month = may,
year = "2020",
address = "Marseille, France",
publisher = "European Language Resources Association",
url = "https://aclanthology.org/2020.lrec-1.428",
pages = "3484--3489",
language = "English",
ISBN = "979-10-95546-34-4",
}
SloTS
@misc{gorenc2022slovene,
title = {Slovene text simplification dataset {SloTS}},
author = {Gorenc, Sabina and Robnik-{\v S}ikonja, Marko},
url = {http://hdl.handle.net/11356/1682},
note = {Slovenian language resource repository {CLARIN}.{SI}},
copyright = {Creative Commons - Attribution 4.0 International ({CC} {BY} 4.0)},
issn = {2820-4042},
year = {2022}
}
Terence and Teacher
@inproceedings{brunato-etal-2015-design,
title = "Design and Annotation of the First {I}talian Corpus for Text Simplification",
author = "Brunato, Dominique and
Dell{'}Orletta, Felice and
Venturi, Giulia and
Montemagni, Simonetta",
booktitle = "Proceedings of the 9th Linguistic Annotation Workshop",
month = jun,
year = "2015",
address = "Denver, Colorado, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/W15-1604",
doi = "10.3115/v1/W15-1604",
pages = "31--41",
}
TextComplexityDE
@article{naderi2019subjective,
title={Subjective Assessment of Text Complexity: A Dataset for German Language},
author={Naderi, Babak and Mohtaj, Salar and Ensikat, Kaspar and M{\"o}ller, Sebastian},
journal={arXiv preprint arXiv:1904.07733},
year={2019}
}
WikiAuto
@inproceedings{acl/JiangMLZX20,
author = {Chao Jiang and
Mounica Maddela and
Wuwei Lan and
Yang Zhong and
Wei Xu},
editor = {Dan Jurafsky and
Joyce Chai and
Natalie Schluter and
Joel R. Tetreault},
title = {Neural {CRF} Model for Sentence Alignment in Text Simplification},
booktitle = {Proceedings of the 58th Annual Meeting of the Association for Computational
Linguistics, {ACL} 2020, Online, July 5-10, 2020},
pages = {7943--7960},
publisher = {Association for Computational Linguistics},
year = {2020},
url = {https://www.aclweb.org/anthology/2020.acl-main.709/}
}
WikiLargeFR
@inproceedings{cardon-grabar-2020-french,
title = "{F}rench Biomedical Text Simplification: When Small and Precise Helps",
author = "Cardon, R{\'e}mi and
Grabar, Natalia",
booktitle = "Proceedings of the 28th International Conference on Computational Linguistics",
month = dec,
year = "2020",
address = "Barcelona, Spain (Online)",
publisher = "International Committee on Computational Linguistics",
url = "https://aclanthology.org/2020.coling-main.62",
doi = "10.18653/v1/2020.coling-main.62",
pages = "710--716",
abstract = "We present experiments on biomedical text simplification in French. We use two kinds of corpora {--} parallel sentences extracted from existing health comparable corpora in French and WikiLarge corpus translated from English to French {--} and a lexicon that associates medical terms with paraphrases. Then, we train neural models on these parallel corpora using different ratios of general and specialized sentences. We evaluate the results with BLEU, SARI and Kandel scores. The results point out that little specialized data helps significantly the simplification.",
}
Data Availability
Public Datasets
Most of the public datasets are available as a part of this MultiSim Repo. A few are still pending availability. For all resources we provide alternative download links.
On Request Datasets
The authors of the original papers must be contacted for on request datasets. Contact information for the authors of each dataset is provided below.
| Dataset | Language | Contact |
|---|---|---|
| CBST | Basque | http://www.ixa.eus/node/13007?language=en itziar.gonzalezd@ehu.eus |
| DSim | Danish | sk@eyejustread.com |
| Newsela EN | English | https://newsela.com/data/ |
| Newsela ES | Spanish | https://newsela.com/data/ |
| German News | German | ebling@cl.uzh.ch |
| Simple German | German | ebling@cl.uzh.ch |
| Simplext | Spanish | horacio.saggion@upf.edu |
| RuAdapt Literature | Russian | Partially Available: https://github.com/Digital-Pushkin-Lab/RuAdapt Full Dataset: anna.dmitrieva@helsinki.fi |
- Downloads last month
- 412