embedding-data/SPECTER

---
license: mit
language:
- en
paperswithcode_id: embedding-data/SPECTER
pretty_name: SPECTER
task_categories:
- sentence-similarity
- paraphrase-mining
task_ids:
- semantic-similarity-classification
---

# Dataset Card for "SPECTER"

## 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 Rationale](#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

- **Homepage:** [https://github.com/allenai/specter](https://github.com/allenai/specter)
- **Repository:** [More Information Needed](https://github.com/allenai/specter/blob/master/README.md)
- **Paper:** [More Information Needed](https://arxiv.org/pdf/2004.07180.pdf)
- **Point of Contact:** [@armancohan](https://github.com/armancohan), [@sergeyf](https://github.com/sergeyf), [@haroldrubio](https://github.com/haroldrubio), [@jinamshah](https://github.com/jinamshah)

### Dataset Summary

Dataset containing triplets (three sentences): anchor, positive, and negative. Contains titles of papers.

Disclaimer: The team releasing SPECTER did not upload the dataset to the Hub and did not write a dataset card.
These steps were done by the Hugging Face team.

## Dataset Structure
Each example in the dataset contains triplets of equivalent sentences and is formatted as a dictionary with the key "set" and a list with the sentences as "value".

Each example is a dictionary with a key, "set", containing a list of three sentences (anchor, positive, and negative):

```
{"set": [anchor, positive, negative]}
{"set": [anchor, positive, negative]}
...
{"set": [anchor, positive, negative]}
```
This dataset is useful for training Sentence Transformers models. Refer to the following post on how to train models using triplets.

### Usage Example
Install the 🤗 Datasets library with `pip install datasets` and load the dataset from the Hub with:
```python
from datasets import load_dataset
dataset = load_dataset("embedding-data/SPECTER")
```
The dataset is loaded as a `DatasetDict` and has the format:
```python
DatasetDict({
train: Dataset({
features: ['set'],
num_rows: 684100
})
})
```
Review an example `i` with:
```python
dataset["train"][i]["set"]
```

### Curation Rationale

[More Information Needed](https://github.com/allenai/specter)

### Source Data

#### Initial Data Collection and Normalization

[More Information Needed](https://github.com/allenai/specter)

#### Who are the source language producers?

[More Information Needed](https://github.com/allenai/specter)

### Annotations

#### Annotation process

[More Information Needed](https://github.com/allenai/specter)

#### Who are the annotators?

[More Information Needed](https://github.com/allenai/specter)

### Personal and Sensitive Information

[More Information Needed](https://github.com/allenai/specter)

## Considerations for Using the Data

### Social Impact of Dataset

[More Information Needed](https://github.com/allenai/specter)

### Discussion of Biases

[More Information Needed](https://github.com/allenai/specter)

### Other Known Limitations

[More Information Needed](https://github.com/allenai/specter)

## Additional Information

### Dataset Curators

[More Information Needed](https://github.com/allenai/specter)

### Licensing Information

[More Information Needed](https://github.com/allenai/specter)

### Citation Information

### Contributions

license: MIT许可证
language:
- 英语
paperswithcode_id: embedding-data/SPECTER
pretty_name: SPECTER
task_categories:
- 句子相似度
- 释义挖掘
task_ids:
- 语义相似度分类

# SPECTER数据集卡片

## 目录
- [数据集描述](#dataset-description)
- [数据集概述](#dataset-summary)
- [支持任务与评测基准](#supported-tasks-and-leaderboards)
- [语言](#languages)
- [数据集结构](#dataset-structure)
- [数据实例](#data-instances)
- [数据字段](#data-fields)
- [数据划分](#data-splits)
- [数据集构建](#dataset-creation)
- [数据集整理依据](#curation-rationale)
- [源数据](#source-data)
- [标注信息](#annotations)
- [个人与敏感信息](#personal-and-sensitive-information)
- [数据集使用注意事项](#considerations-for-using-the-data)
- [数据集的社会影响](#social-impact-of-dataset)
- [偏差讨论](#discussion-of-biases)
- [其他已知局限性](#other-known-limitations)
- [附加信息](#additional-information)
- [数据集整理者](#dataset-curators)
- [许可证信息](#licensing-information)
- [引用信息](#citation-information)
- [贡献信息](#contributions)

## 数据集描述
- **主页：** [https://github.com/allenai/specter](https://github.com/allenai/specter)
- **代码仓库：** [需补充更多信息](https://github.com/allenai/specter/blob/master/README.md)
- **相关论文：** [需补充更多信息](https://arxiv.org/pdf/2004.07180.pdf)
- **联系方式：** [@armancohan](https://github.com/armancohan)、[@sergeyf](https://github.com/sergeyf)、[@haroldrubio](https://github.com/haroldrubio)、[@jinamshah](https://github.com/jinamshah)

### 数据集概述
本数据集包含三元组结构（即三个句子）：锚点句（anchor）、正样本句（positive）与负样本句（negative），所有句子均为学术论文的标题。
免责声明：SPECTER的开发团队并未将该数据集上传至Hugging Face Hub，也未编写本数据集卡片，相关上传与卡片编写工作均由Hugging Face团队完成。

## 数据集结构
数据集中的每个样本均由等价句子三元组构成，格式为以"set"为键、句子列表为值的字典。
每个样本均为包含"set"键的字典，其对应值为包含三个句子（锚点句anchor、正样本句positive、负样本句negative）的列表：

{"set": [anchor, positive, negative]}
{"set": [anchor, positive, negative]}
...
{"set": [anchor, positive, negative]}

本数据集可用于训练Sentence Transformers模型，可参考以下教程学习如何使用三元组训练模型。

### 使用示例
通过`pip install datasets`安装🤗 Datasets库，并使用以下代码从Hub加载数据集：
python
from datasets import load_dataset
dataset = load_dataset("embedding-data/SPECTER")

加载后的数据集为`DatasetDict`格式，具体结构如下：
python
DatasetDict({
train: Dataset({
features: ['set'],
num_rows: 684100
})
})

可通过以下代码查看第`i`个样本：
python
dataset["train"][i]["set"]

### 数据集整理依据
[需补充更多信息](https://github.com/allenai/specter)

### 源数据

#### 初始数据收集与标准化
[需补充更多信息](https://github.com/allenai/specter)

#### 源语言生产者是谁？
[需补充更多信息](https://github.com/allenai/specter)

### 标注信息

#### 标注流程
[需补充更多信息](https://github.com/allenai/specter)

#### 标注者是谁？
[需补充更多信息](https://github.com/allenai/specter)

### 个人与敏感信息
[需补充更多信息](https://github.com/allenai/specter)

## 数据集使用注意事项

### 数据集的社会影响
[需补充更多信息](https://github.com/allenai/specter)

### 偏差讨论
[需补充更多信息](https://github.com/allenai/specter)

### 其他已知局限性
[需补充更多信息](https://github.com/allenai/specter)

## 附加信息

### 数据集整理者
[需补充更多信息](https://github.com/allenai/specter)

### 许可证信息
[需补充更多信息](https://github.com/allenai/specter)

### 引用信息

### 贡献信息

原始信息汇总

数据集概述：SPECTER

数据集描述

数据集总结

• 内容: 包含三元组（三个句子）：锚点句、正例句和负例句，以及论文标题。
• 格式: 每个示例为字典格式，键为"set"，值为包含三个句子的列表。

支持的任务和排行榜

• 任务类别: 句子相似度、释义挖掘
• 任务ID: 语义相似度分类

语言

• 支持语言: 英语

数据集结构

数据实例

• 结构: 每个数据实例为一个字典，包含键"set"和值为一个包含三个句子的列表。

• 示例:

  {"set": [anchor, positive, negative]} {"set": [anchor, positive, negative]} ... {"set": [anchor, positive, negative]}

数据字段

• 字段: "set"，包含三个句子（锚点句、正例句、负例句）的列表。

数据分割

• 示例: python DatasetDict({ train: Dataset({ features: [set], num_rows: 684100 }) })

使用示例

• 加载数据集: python from datasets import load_dataset dataset = load_dataset("embedding-data/SPECTER")

• 查看示例: python dataset["train"][i]["set"]

许可证信息

• 许可证: MIT

搜集汇总

背景与挑战

背景概述

SPECTER是一个用于训练Sentence Transformers模型的英文句子三元组数据集，包含锚点、正例和负例句子，主要用于句子相似性和释义挖掘任务。该数据集以字典格式存储，训练集规模为684100个三元组示例。