Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Languages:
Japanese
Size:
10K - 100K
License:
dataset_info: | |
features: | |
- name: premise | |
dtype: large_string | |
- name: hypothesis | |
dtype: large_string | |
- name: label | |
dtype: | |
class_label: | |
names: | |
'0': entailment | |
'1': neutral | |
'2': contradiction | |
splits: | |
- name: train | |
num_bytes: 2025224 | |
num_examples: 12667 | |
- name: validation | |
num_bytes: 256996 | |
num_examples: 1583 | |
- name: test | |
num_bytes: 250810 | |
num_examples: 1584 | |
download_size: 1150200 | |
dataset_size: 2533030 | |
configs: | |
- config_name: default | |
data_files: | |
- split: train | |
path: data/train-* | |
- split: validation | |
path: data/validation-* | |
- split: test | |
path: data/test-* | |
license: bsd-3-clause | |
task_categories: | |
- text-classification | |
language: | |
- ja | |
tags: | |
- evaluation | |
- benchmark | |
- nli | |
pretty_name: 'JSeM: Japanese semantic test suite' | |
# JSeM: Japanese semantic test suite (Japanese FraCaS and extensions) | |
叙述文間の含意関係は、言語学においては意味論の中心的な説明対象の一つであるとともに、理論を検証するためのベンチマークとして用いられています。 | |
また近年の自然言語処理においては、含意関係認識(Recognizing Textual Entailment: RTE)が意味処理タスクの中核となっています。 | |
- 前提(premise):一つの文 | |
- 仮説(hypothesis):一つの文 | |
- 判定(answer):1.と2.の間に含意関係があるかどうかについての母語話者の判断(`entailment`, `neutral`あるいは`contradiction`) | |
このテストセットでは、FraCaS test suite(Cooper et al.(1996)で公開されたオリジナルのテストセット, | |
およびBill MacCartney氏による同セットのXML版)の方針にならい、言語現象ごとに含意関係のテストをまとめています。 | |
FraCaSで扱われている現象については、FraCaSのテスト項目へのリンクの付いたデータを作成しています(下記のβ版を参照)。 | |
また、FraCaSで扱われていない現象や、日本語独自の現象の関わるテストも多く構築しており、今後も追加する予定です。 | |
## Dataset Details | |
### Dataset Description | |
- **Created by:** DaisukeBekki | |
- **Language(s) (NLP):** Japanese | |
- **License:** BSD 3-Clause | |
### Dataset Sources | |
- **Repository:** [DaisukeBekki/JSeM](https://github.com/DaisukeBekki/JSeM) | |
- **Paper:** [More Information Needed] | |
## Citation | |
<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. --> | |
**BibTeX:** | |
[More Information Needed] | |
**APA:** | |
[More Information Needed] | |