jpwahle/autoencoder-paraphrase-dataset

---
annotations_creators:
- machine-generated
language:
- en
language_creators:
- machine-generated
license:
- cc-by-4.0
multilinguality:
- monolingual
pretty_name: Autoencoder Paraphrase Dataset (BERT, RoBERTa, Longformer)
size_categories:
- 100K<n<1M
source_datasets:
- original
tags:
- bert
- roberta
- longformer
- plagiarism
- paraphrase
- academic integrity
- arxiv
- wikipedia
- theses
task_categories:
- text-classification
- text-generation
task_ids: []
paperswithcode_id: are-neural-language-models-good-plagiarists-a
dataset_info:
- split: train
download_size: 2980464
dataset_size: 2980464
- split: test
download_size: 1690032
dataset_size: 1690032
---

# Dataset Card for Machine Paraphrase Dataset (MPC)

## Table of Contents
- [Table of Contents](#table-of-contents)
- [Dataset Description](#dataset-description)
- [Dataset Summary](#dataset-summary)
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
- [Languages](#languages)
- [Dataset Structure](#dataset-structure)
- [Data Instances](#data-instances)
- [Data Fields](#data-fields)
- [Data Splits](#data-splits)
- [Dataset Creation](#dataset-creation)
- [Curation Rat1.ionale](#curation-rationale)
- [Source Data](#source-data)
- [Annotations](#annotations)
- [Personal and Sensitive Information](#personal-and-sensitive-information)
- [Considerations for Using the Data](#considerations-for-using-the-data)
- [Social Impact of Dataset](#social-impact-of-dataset)
- [Discussion of Biases](#discussion-of-biases)
- [Other Known Limitations](#other-known-limitations)
- [Additional Information](#additional-information)
- [Dataset Curators](#dataset-curators)
- [Licensing Information](#licensing-information)
- [Citation Information](#citation-information)
- [Contributions](#contributions)

## Dataset Description

- **Paper:** https://ieeexplore.ieee.org/document/9651895
- **Total size:** 2.23 GB
- **Train size:** 1.52 GB
- **Test size:** 861 MB

### Dataset Summary

The Autoencoder Paraphrase Corpus (APC) consists of ~200k examples of original, and paraphrases using three neural language models.
It uses three models (BERT, RoBERTa, Longformer) on three source texts (Wikipedia, arXiv, student theses).
The examples are aligned, i.e., we sample the same paragraphs for originals and paraphrased versions.

### How to use it
You can load the dataset using the `load_dataset` function:

```python
from datasets import load_dataset

ds = load_dataset("jpwahle/autoencoder-paraphrase-dataset")
print(ds[0])

#OUTPUT:
{
'text': 'War memorial formally unveiled on Whit Monday 16 May 1921 by the Prince of Wales later King Edward VIII with Lutyens in attendance At the unveiling ceremony Captain Fortescue gave a speech during wherein he announced that 11 600 men and women from Devon had been inval while serving in imperialist war He later stated that some 63 700 8 000 regulars 36 700 volunteers 19 000 conscripts had served in the armed forces The heroism of the dead are recorded on a roll of honour of which three copies were made one for Exeter Cathedral one To be held by Tasman county council and another honoring the Prince of Wales placed in a hollow in bedrock base of the war memorial The princes visit generated considerable excitement in the area Thousands of spectators lined the street to greet his motorcade and shops on Market High Street hung out banners with welcoming messages After the unveiling Edward spent ten days touring the local area',
'label': 1,
'dataset': 'wikipedia',
'method': 'longformer'
}
```

### Supported Tasks and Leaderboards

Paraphrase Identification

### Languages

English

## Dataset Structure

### Data Instances

```json
{
'text': 'War memorial formally unveiled on Whit Monday 16 May 1921 by the Prince of Wales later King Edward VIII with Lutyens in attendance At the unveiling ceremony Captain Fortescue gave a speech during wherein he announced that 11 600 men and women from Devon had been inval while serving in imperialist war He later stated that some 63 700 8 000 regulars 36 700 volunteers 19 000 conscripts had served in the armed forces The heroism of the dead are recorded on a roll of honour of which three copies were made one for Exeter Cathedral one To be held by Tasman county council and another honoring the Prince of Wales placed in a hollow in bedrock base of the war memorial The princes visit generated considerable excitement in the area Thousands of spectators lined the street to greet his motorcade and shops on Market High Street hung out banners with welcoming messages After the unveiling Edward spent ten days touring the local area',
'label': 1,
'dataset': 'wikipedia',
'method': 'longformer'
}
```

### Data Fields

| Feature | Description |
| --- | --- |
| `text` | The unique identifier of the paper. |
| `label` | Whether it is a paraphrase (1) or the original (0). |
| `dataset` | The source dataset (Wikipedia, arXiv, or theses). |
| `method` | The method used (bert, roberta, longformer). |

### Data Splits

- train (Wikipedia x [bert, roberta, longformer])
- test ([Wikipedia, arXiv, theses] x [bert, roberta, longformer])

## Dataset Creation

### Curation Rationale

Providing a resource for testing against autoencoder paraprhased plagiarism.

### Source Data

#### Initial Data Collection and Normalization

- Paragraphs from `featured articles` from the English Wikipedia dump
- Paragraphs from full-text pdfs of arXMLiv
- Paragraphs from full-text pdfs of Czech student thesis (bachelor, master, PhD).

#### Who are the source language producers?

[More Information Needed]

### Annotations

#### Annotation process

[More Information Needed]

#### Who are the annotators?

[More Information Needed]

### Personal and Sensitive Information

[More Information Needed]

## Considerations for Using the Data

### Social Impact of Dataset

[More Information Needed]

### Discussion of Biases

[More Information Needed]

### Other Known Limitations

[More Information Needed]

## Additional Information

### Dataset Curators

[Jan Philip Wahle](https://jpwahle.com/)

### Licensing Information

The Autoencoder Paraphrase Dataset is released under CC BY-NC 4.0. By using this corpus, you agree to its usage terms.

### Citation Information

```bib
@inproceedings{9651895,
title = {Are Neural Language Models Good Plagiarists? A Benchmark for Neural Paraphrase Detection},
author = {Wahle, Jan Philip and Ruas, Terry and Meuschke, Norman and Gipp, Bela},
year = 2021,
booktitle = {2021 ACM/IEEE Joint Conference on Digital Libraries (JCDL)},
volume = {},
number = {},
pages = {226--229},
doi = {10.1109/JCDL52503.2021.00065}
}
```

### Contributions

Thanks to [@jpwahle](https://github.com/jpwahle) for adding this dataset.

原始信息汇总

数据集概述

基本信息

• 数据集名称: Autoencoder Paraphrase Dataset (BERT, RoBERTa, Longformer)
• 语言: 英语
• 许可证: CC BY-4.0
• 数据集大小: 100K<n<1M
• 多语言性: 单语
• 标签: bert, roberta, longformer, plagiarism, paraphrase, academic integrity, arxiv, wikipedia, theses
• 任务类别: 文本分类, 文本生成

数据集描述

数据集摘要

Autoencoder Paraphrase Corpus (APC) 包含约200k个原始文本和使用三种神经语言模型生成的释义示例。使用了三种模型（BERT, RoBERTa, Longformer）和三种源文本（Wikipedia, arXiv, 学生论文）。示例是对齐的，即从原始文本和释义版本中采样相同的段落。

支持的任务和排行榜

• 释义识别

语言

• 英语

数据集结构

数据实例

json { text: War memorial formally unveiled on Whit Monday 16 May 1921 by the Prince of Wales later King Edward VIII with Lutyens in attendance At the unveiling ceremony Captain Fortescue gave a speech during wherein he announced that 11 600 men and women from Devon had been inval while serving in imperialist war He later stated that some 63 700 8 000 regulars 36 700 volunteers 19 000 conscripts had served in the armed forces The heroism of the dead are recorded on a roll of honour of which three copies were made one for Exeter Cathedral one To be held by Tasman county council and another honoring the Prince of Wales placed in a hollow in bedrock base of the war memorial The princes visit generated considerable excitement in the area Thousands of spectators lined the street to greet his motorcade and shops on Market High Street hung out banners with welcoming messages After the unveiling Edward spent ten days touring the local area, label: 1, dataset: wikipedia, method: longformer }

数据字段

特征	描述
text	文本内容
label	是否为释义（1）或原始文本（0）
dataset	源数据集（Wikipedia, arXiv, 论文）
method	使用的方法（bert, roberta, longformer）

数据分割

• 训练集 (Wikipedia x [bert, roberta, longformer])
• 测试集 ([Wikipedia, arXiv, 论文] x [bert, roberta, longformer])

数据集创建

策划理由

提供一个用于测试自动编码器释义抄袭的资源。

源数据

初始数据收集和规范化

• 来自英文维基百科转储的“特色文章”段落
• 来自arXMLiv全文pdf的段落
• 来自捷克学生论文（学士、硕士、博士）全文pdf的段落

使用数据集的注意事项

数据集的社会影响

[更多信息需要]

偏见讨论

[更多信息需要]

其他已知限制

[更多信息需要]

附加信息

数据集策展人

• Jan Philip Wahle

许可信息

Autoencoder Paraphrase Dataset 在 CC BY-NC 4.0 许可下发布。使用此语料库即表示您同意其使用条款。

引用信息

bib @inproceedings{9651895, title = {Are Neural Language Models Good Plagiarists? A Benchmark for Neural Paraphrase Detection}, author = {Wahle, Jan Philip and Ruas, Terry and Meuschke, Norman and Gipp, Bela}, year = 2021, booktitle = {2021 ACM/IEEE Joint Conference on Digital Libraries (JCDL)}, volume = {}, number = {}, pages = {226--229}, doi = {10.1109/JCDL52503.2021.00065} }

贡献

感谢 @jpwahle 添加此数据集。