sxu/CANLI

---
license: afl-3.0

annotations_creators:
- expert-generated
language:
- cn
language_creators:
- expert-generated
multilinguality:
- monolingual
size_categories:
- 1K<n<10K
---

# Dataset Card for CANLI

### Dataset Summary
[CANLI: The Chinese Causative-Passive Homonymy Disambiguation: an Adversarial Dataset for NLI and a Probing Task](http://www.lrec-conf.org/proceedings/lrec2022/pdf/2022.lrec-1.460.pdf)

The disambiguation of causative-passive homonymy (CPH) is potentially tricky for machines, as the causative and the passive
are not distinguished by the sentences syntactic structure. By transforming CPH disambiguation to a challenging natural
language inference (NLI) task, we present the first Chinese Adversarial NLI challenge set (CANLI). We show that the pretrained
transformer model RoBERTa, fine-tuned on an existing large-scale Chinese NLI benchmark dataset, performs poorly on CANLI.
We also employ Word Sense Disambiguation as a probing task to investigate to what extent the CPH feature is captured in
the models internal representation. We find that the models performance on CANLI does not correspond to its internal
representation of CPH, which is the crucial linguistic ability central to the CANLI dataset.

### Languages
Chinese Mandarin

# Citation Information

@inproceedings{xu-markert-2022-chinese,
title = "The {C}hinese Causative-Passive Homonymy Disambiguation: an adversarial Dataset for {NLI} and a Probing Task",
author = "Xu, Shanshan and Markert, Katja",
booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
month = jun,
year = "2022",
address = "Marseille, France",
publisher = "European Language Resources Association",
url = "https://aclanthology.org/2022.lrec-1.460",
pages = "4316--4323",
}

原始信息汇总

数据集概述

基本信息

• 数据集名称: CANLI
• 数据集全称: The Chinese Causative-Passive Homonymy Disambiguation: an Adversarial Dataset for NLI and a Probing Task
• 语言: 中文（Chinese Mandarin）
• 许可证: afl-3.0
• 多语言性: 单语（monolingual）
• 大小: 1K<n<10K

数据集描述

• 目标: 解决中文中因果-被动同音异义词的歧义问题。
• 方法: 将因果-被动同音异义词的歧义问题转化为自然语言推理（NLI）任务。
• 发现: 预训练的转换器模型RoBERTa，在现有的中文NLI基准数据集上进行微调后，在CANLI数据集上的表现不佳。
• 应用: 通过词义消歧作为探测任务，研究模型内部表示中因果-被动同音异义特征的捕捉程度。

引用信息

@inproceedings{xu-markert-2022-chinese, title = "The {C}hinese Causative-Passive Homonymy Disambiguation: an adversarial Dataset for {NLI} and a Probing Task", author = "Xu, Shanshan and Markert, Katja", booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference", month = jun, year = "2022", address = "Marseille, France", publisher = "European Language Resources Association", url = "https://aclanthology.org/2022.lrec-1.460", pages = "4316--4323", }