metadata

annotations_creators:
  - expert-generated
  - crowdsourced
language:
  - en
language_creators:
  - found
  - crowdsourced
  - expert-generated
license:
  - c-uda
multilinguality:
  - monolingual
paperswithcode_id: totto
pretty_name: HiTab
size_categories:
  - 10K<n<100K
source_datasets:
  - original
  - extended|totto
tags: []
task_categories:
  - table-to-text
  - question-answering
task_ids: []
dataset_info:
  features:
    - name: id
      dtype: string
    - name: table_id
      dtype: string
    - name: table_source
      dtype: string
    - name: sentence_id
      dtype: string
    - name: sub_sentence_id
      dtype: string
    - name: sub_sentence
      dtype: string
    - name: question
      dtype: string
    - name: answer
      dtype: large_string
    - name: aggregation
      dtype: large_string
    - name: linked_cells
      dtype: large_string
    - name: answer_formulas
      dtype: large_string
    - name: reference_cells_map
      dtype: large_string
    - name: table_content
      dtype: large_string
  splits:
    - name: train
      num_bytes: 36419103
      num_examples: 7417
    - name: validation
      num_bytes: 8312699
      num_examples: 1671
    - name: test
      num_bytes: 7710891
      num_examples: 1584
  download_size: 6462957
  dataset_size: 52442693

Dataset Card for HiTab

Dataset Description

Homepage: https://www.microsoft.com/en-us/research/publication/hitab-a-hierarchical-table-dataset-for-question-answering-and-natural-language-generation/
Repository: https://github.com/microsoft/HiTab/blob/main/LICENSE
Paper: https://aclanthology.org/2022.acl-long.78/

Dataset Summary

HiTab is a dataset for question answering and data-to-text over hierarchical tables . It contains 10,672 samples and 3,597 tables from statistical reports (StatCan, NSF) and Wikipedia (ToTTo). 98.1% of the tables in HiTab are with hierarchies.

Supported Tasks and Leaderboards

Table-to-Text Generation, Question Answering

Languages

English

Data Instances

3,597 tables, 10,686 sentences

Data Splits

train, test, validation

Dataset Creation

During the dataset annotation process, annotators first manually collect tables and descriptive sentences highly-related to tables on statistical websites written by professional analysts. Then these descriptions are revised to questions to preserve the original meanings and analyses.

Licensing Information

This dataset follows the Computational Use of Data Agreement v1.0.

Citation Information

@article{cheng2021hitab,
  title={HiTab: A Hierarchical Table Dataset for Question Answering and Natural Language Generation},
  author={Cheng, Zhoujun and Dong, Haoyu and Wang, Zhiruo and Jia, Ran and Guo, Jiaqi and Gao, Yan and Han, Shi and Lou, Jian-Guang and Zhang, Dongmei},
  journal={arXiv preprint arXiv:2108.06712},
  year={2021}
}