clarin-pl/PUGG

---
annotations_creators:
- expert-generated
language_creators: []
language:
- pl
license:
- cc-by-sa-4.0
multilinguality:
- monolingual
size_categories:
- 1K<n<10K
- 10K<n<100K
source_datasets:
- original
task_categories:
- question-answering
- text-retrieval
task_ids:
- extractive-qa
- document-retrieval
pretty_name: 'PUGG: KBQA, MRC, IR dataset for Polish'
tags:
- knowledge graph
- KBQA
- wikipedia
- wikidata
configs:
- config_name: kbqa_all
data_files:
- split: train
path: kbqa/*/train.jsonl
- split: test
path: kbqa/*/test.jsonl
- config_name: kbqa_natural
data_files:
- split: train
path: kbqa/natural/train.jsonl
- split: test
path: kbqa/natural/test.jsonl
- config_name: kbqa_template-based
data_files:
- split: train
path: kbqa/template-based/train.jsonl
- split: test
path: kbqa/template-based/test.jsonl
- config_name: mrc
data_files:
- split: train
path: mrc/train.jsonl
- split: test
path: mrc/test.jsonl
- config_name: ir_corpus
data_files:
- split: test
path: ir/corpus.jsonl
- config_name: ir_queries
data_files:
- split: test
path: ir/queries.jsonl
- config_name: ir_qrels
data_files:
- split: test
path: ir/qrels/test.jsonl
---
# PUGG: KBQA, MRC, IR Dataset for Polish

## Description

This repository contains the PUGG dataset designed for three NLP tasks in the Polish language:

- KBQA (Knowledge Base Question Answering)
- MRC (Machine Reading Comprehension)
- IR (Information Retrieval)

## Paper

For more detailed information, please refer to our research paper titled:

**"Developing PUGG for Polish: A Modern Approach to KBQA, MRC, and IR Dataset Construction"**

Authored by:
* Albert Sawczyn
* Katsiaryna Viarenich
* Konrad Wojtasik
* Aleksandra Domogała
* Marcin Oleksy
* Maciej Piasecki
* Tomasz Kajdanowicz

**The paper was accepted for ACL 2024 (findings).**

## Repositories

The dataset is available in the following repositories:

* [General](https://huggingface.co/datasets/clarin-pl/PUGG) **(this repository)** - contains all tasks (KBQA, MRC, IR*)

For more straightforward usage, the tasks are also available in separate repositories:

* [KBQA](https://huggingface.co/datasets/clarin-pl/PUGG_KBQA)
* [MRC](https://huggingface.co/datasets/clarin-pl/PUGG_MRC)
* [IR](https://huggingface.co/datasets/clarin-pl/PUGG_IR)

The knowledge graph for KBQA task is available in the following repository:

* [Knowledge Graph](https://huggingface.co/datasets/clarin-pl/PUGG_KG)

Note: If you want to utilize the IR task in the BEIR format (`qrels` in `.tsv` format), please
download the [IR](https://huggingface.co/datasets/clarin-pl/PUGG_IR) repository.

## Links

* Code:
* [Github](https://github.com/CLARIN-PL/PUGG)
* Paper:
* ACL - TBA
* [Arxiv](https://arxiv.org/abs/2408.02337)

## Citation

```bibtex
@misc{sawczyn2024developingpuggpolishmodern,
title={Developing PUGG for Polish: A Modern Approach to KBQA, MRC, and IR Dataset Construction},
author={Albert Sawczyn and Katsiaryna Viarenich and Konrad Wojtasik and Aleksandra Domogała and Marcin Oleksy and Maciej Piasecki and Tomasz Kajdanowicz},
year={2024},
eprint={2408.02337},
archivePrefix={arXiv},
primaryClass={cs.AI},
url={https://arxiv.org/abs/2408.02337},
}
```

## Contact

[e-mail]

## Usage

```python
from datasets import load_dataset

# loading KBQA (all)
dataset = load_dataset("clarin-pl/PUGG", "kbqa_all")
print(dataset)

# loading KBQA (natural)
dataset = load_dataset("clarin-pl/PUGG", "kbqa_natural")
print(dataset)

# loading KBQA (template-based)
dataset = load_dataset("clarin-pl/PUGG", "kbqa_template-based")
print(dataset)

# loading MRC

dataset = load_dataset("clarin-pl/PUGG", "mrc")
print(dataset)

# loading IR

## corpus
dataset = load_dataset("clarin-pl/PUGG", "ir_corpus")
print(dataset)

## queries
dataset = load_dataset("clarin-pl/PUGG", "ir_queries")
print(dataset)

## qrels
dataset = load_dataset("clarin-pl/PUGG", "ir_qrels")
print(dataset)
```

原始信息汇总

PUGG: KBQA, MRC, IR 数据集概述

数据集描述

本数据集名为PUGG，是根据研究论文《Developing PUGG for Polish: A Modern Approach to KBQA, MRC, and IR Dataset Construction》创建的，该论文由以下作者撰写：

• Albert Sawczyn
• Katsiaryna Viarenich
• Konrad Wojtasik
• Aleksandra Domogała
• Marcin Oleksy
• Maciej Piasecki
• Tomasz Kajdanowicz

该论文已被ACL 2024（Findings）接受。

数据集用途

PUGG数据集旨在支持波兰语的KBQA（知识库问答）、MRC（机器阅读理解）和IR（信息检索）研究。