luheng/qa_srl

Dataset Card for QA-SRL

Dataset Summary

we model predicate-argument structure of a sentence with a set of question-answer pairs. our method allows practical large-scale annotation of training data. We focus on semantic rather than syntactic annotation, and introduce a scalable method for gathering data that allows both training and evaluation.

Supported Tasks and Leaderboards

[More Information Needed]

Languages

This dataset is in english language.

Dataset Structure

Data Instances

We use question-answer pairs to model verbal predicate-argument structure. The questions start with wh-words (Who, What, Where, What, etc.) and contains a verb predicate in the sentence; the answers are phrases in the sentence. For example:

UCD finished the 2006 championship as Dublin champions , by beating St Vincents in the final .

Predicate	Question	Answer
Finished	Who finished something?	UCD
Finished	What did someone finish?	the 2006 championship
Finished	What did someone finish something as?	Dublin champions
Finished	How did someone finish something?	by beating St Vincents in the final
beating	Who beat someone?	UCD
beating	When did someone beat someone?	in the final
beating	Who did someone beat?	St Vincents

Data Fields

Annotations provided are as follows:

• sentence: contains tokenized sentence
• sent_id: is the sentence identifier
• predicate_idx:the index of the predicate (its position in the sentence)
• predicate: the predicate token
• question: contains the question which is a list of tokens. The question always consists of seven slots, as defined in the paper. The empty slots are represented with a marker “_”. The question ends with question mark.
• answer: list of answers to the question

Data Splits

Dataset	Sentences	Verbs	QAs
newswire-train	744	2020	4904
newswire-dev	249	664	1606
newswire-test	248	652	1599
Wikipedia-train	1174	2647	6414
Wikipedia-dev	392	895	2183
Wikipedia-test	393	898	2201

Please note This dataset only has wikipedia data. Newswire dataset needs CoNLL-2009 English training data to get the complete data. This training data is under license. Thus, newswire dataset is not included in this data.

Dataset Creation

Curation Rationale

[More Information Needed]

Source Data

Initial Data Collection and Normalization

We annotated over 3000 sentences (nearly 8,000 verbs) in total across two domains: newswire (PropBank) and Wikipedia.

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

non-expert annotators were given a short tutorial and a small set of sample annotations (about 10 sentences). Annotators were hired if they showed good understanding of English and the task. The entire screening process usually took less than 2 hours.

Who are the annotators?

10 part-time, non-exper annotators from Upwork (Previously oDesk)

Personal and Sensitive Information

[More Information Needed]

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

Luheng He

Licensing Information

[More Information Needed]

Citation Information

@InProceedings{huggingface:dataset,
title = {QA-SRL: Question-Answer Driven Semantic Role Labeling},
authors={Luheng He, Mike Lewis, Luke Zettlemoyer},
year={2015}
publisher = {cs.washington.edu},
howpublished={\\url{https://dada.cs.washington.edu/qasrl/#page-top}},
}

Contributions

Thanks to @bpatidar for adding this dataset.