Datasets:
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
google-research-datasets/google_wellformed_query(複製)
データの内容
英語のクエリが整った自然な質問の形になっているかを判定するためのデータ。1 行が 1 件のクエリで、文を content、整形されている確率を rating に持つ。Paralex コーパスから選んだクエリに、5 人のクラウドワーカーが整形されているかを 2 値で判定し、その平均を rating にしている。ライセンスは CC BY-SA 4.0 である。
データの概要
| 項目 | 値 |
|---|---|
| 行数 | 25,100(1 構成・3 分割の合計) |
| 列数 | 2 |
| 容量 | 9.82 KB(ファイル 3 本) |
| ファイルの形式 | py 1 |
| 言語 | en |
| 公開 | public |
構成と分割
| 構成(config) | 分割(split)ごとの行数 | 列数 |
|---|---|---|
default |
train 17,500、test 3,850、validation 3,750 |
2 |
列(構成 default)
| 列 | 型 |
|---|---|
rating |
float32 |
content |
string |
行数と列は元のカードの YAML(dataset_info)の記載。データセットビューアでは読めなかった。全データセットの一覧は HYUKI5/dataset-catalog にある。
複製について
google-research-datasets/google_wellformed_query をリビジョン 287f895e19da0c028407dd0b6356efffc9401ecf の時点でそのまま複製したもの。中身は変えていない。判断モデル Laya(convaiinnovations/laya)の微調整に使う候補として、調査メモ(Laya 学習用データセット調査)に載せたデータを集めている。
出典
- 元のデータセット: https://huggingface.co/datasets/google-research-datasets/google_wellformed_query
- 複製したリビジョン:
287f895e19da0c028407dd0b6356efffc9401ecf(2024-01-18 11:04:23+00:00) - 元のライセンス表記:
['cc-by-sa-4.0']
論文
アノテーションの方法
収集の根拠
クエリを理解することは難しい問題である。ユーザーは「単語の寄せ集め」(word salad)のようなテキストを日常的に入力するからである。しかし、クエリが整った質問に似ていれば、自然言語処理のパイプラインはより正確に解釈でき、後段の誤りの連鎖を減らせる。したがって、クエリが整形されているかどうかを判定すれば、クエリの理解を高められる。このデータセットは、整った自然言語の質問を判定するという新しいタスクを導入する。
元データ
Paralex コーパス(Fader et al., 2013)を使った。このコーパスには、雑音を含む言い換えの質問のペアが含まれる。これらの質問は WikiAnswers(質問と回答のフォーラム)でユーザーが入力したもので、ウェブ検索のクエリに近い形("5 parts of chloroplast?")と、整った質問("What is the punishment for grand theft?")の両方からなる。
初期のデータ収集と正規化
コーパスから抽出した一意なクエリのリストから、選ばれたクエリのどの 2 つも互いに言い換えにならないように、25,100 件のクエリを選んだ。
アノテーションの手順
クエリは、次のすべてを満たすかどうかに応じて、整形された質問か、整形されていない質問かに分類される:
- クエリが文法的である。
- クエリが明示的な質問である。
- クエリにスペルミスがない。
微調整に使うときの注意
- 調査メモの台帳での区分: B(継承条件・gated・モデル由来の条件・上流の規約などが付く)。区分は記載の文言で機械的に付けたもので、法的な判定ではない。学習に入れる前に元のカードとライセンスの原文で確かめる。
- Laya の入力は 2,048 token まで(ModernBERT-large / mmBERT)。長い本文は先頭から黙って切られるので、切れる行は事前に除くか、要約・切り出しをしてから入れる。
- 分割は記事・文書単位で行い、同じ本文から作った別の問いを学習側と較正側に分けない。温度と閾値は学習に使っていない分割で別々に当てる。
調査メモの記述
| メモ | 節 | ライセンス(記載) | 確認 | 注記 |
|---|---|---|---|---|
jev_sources_tasksource_bekko.csv |
CSV(tasksource-jev-typed-decisions) | cc-by-sa-4.0 / license_use=commercial | 確認済 | license_use=commercial / [Jev 元データ source] human_votes_aggregated_to_soft_target(source annotation; agg |
元のカード(日本語訳)
以下は元のデータセットのカードの本文を日本語に訳したもの(機械翻訳)。訳に疑問があるときは、末尾の原文を確かめる。
Google Query-wellformedness データセットのカード
目次
- データセットの説明
- データセットの概要
- 対応するタスクとリーダーボード
- 言語
- データセットの構造
- データのインスタンス
- データのフィールド
- データ分割
- データセットの作成
- 収集の根拠
- 元データ
- アノテーション
- 個人情報と機微な情報
- データを使う際の考慮事項
- データセットの社会的影響
- バイアスの議論
- その他の既知の制限
- 追加情報
- データセットの作成者
- ライセンス情報
- 引用情報
- 貢献者
データセットの説明
データセットの概要
Google の query wellformedness データセットは、Paralex コーパスの 25,100 件のクエリに対して、クラウドソーシングで整形性(well-formedness)のアノテーションを付けて作られた。各クエリは 5 人の評価者がそれぞれ 1 または 0 で、クエリが整形されているかどうかを評価した。
対応するタスクとリーダーボード
[More Information Needed]
言語
英語
データセットの構造
データのインスタンス
{'rating': 0.2, 'content': 'The European Union includes how many ?'}
データのフィールド
rating: 0〜1 のfloatsentence: 評価したいクエリ
データ分割
| Train | Valid | Test | |
|---|---|---|---|
| 入力文 | 17500 | 3750 | 3850 |
データセットの作成
収集の根拠
クエリを理解することは難しい問題である。ユーザーは「単語の寄せ集め」(word salad)のようなテキストを日常的に入力するからである。しかし、クエリが整った質問に似ていれば、自然言語処理のパイプラインはより正確に解釈でき、後段の誤りの連鎖を減らせる。したがって、クエリが整形されているかどうかを判定すれば、クエリの理解を高められる。このデータセットは、整った自然言語の質問を判定するという新しいタスクを導入する。
元データ
Paralex コーパス(Fader et al., 2013)を使った。このコーパスには、雑音を含む言い換えの質問のペアが含まれる。これらの質問は WikiAnswers(質問と回答のフォーラム)でユーザーが入力したもので、ウェブ検索のクエリに近い形("5 parts of chloroplast?")と、整った質問("What is the punishment for grand theft?")の両方からなる。
初期のデータ収集と正規化
コーパスから抽出した一意なクエリのリストから、選ばれたクエリのどの 2 つも互いに言い換えにならないように、25,100 件のクエリを選んだ。
元データの作成者は誰か?
[More Information Needed]
アノテーション
アノテーションの手順
クエリは、次のすべてを満たすかどうかに応じて、整形された質問か、整形されていない質問かに分類される:
- クエリが文法的である。
- クエリが明示的な質問である。
- クエリにスペルミスがない。
アノテーターは誰か?
各クエリは 5 人の異なるクラウドワーカーにより、整形されているかどうかを示す 2 値のラベルが付けられた。クエリが整形されている確率として、5 人のアノテーターの評価の平均を報告した。
個人情報と機微な情報
[More Information Needed]
データを使う際の考慮事項
データセットの社会的影響
[More Information Needed]
バイアスの議論
[More Information Needed]
その他の既知の制限
[More Information Needed]
追加情報
データセットの作成者
[More Information Needed]
ライセンス情報
Query-wellformedness データセットは CC BY-SA 4.0 のライセンスで提供される。第三者のコンテンツやデータは、明示または黙示を問わず、いかなる保証もなく「現状のまま」提供される。
引用情報
@InProceedings{FaruquiDas2018,
title = {{Identifying Well-formed Natural Language Questions}},
author = {Faruqui, Manaal and Das, Dipanjan},
booktitle = {Proc. of EMNLP},
year = {2018}
}
貢献者
このデータセットを追加してくれた @vasudevgupta7 に感謝する。
原文(英語)
Dataset Card for Google Query-wellformedness Dataset
Dataset Summary
Google's query wellformedness dataset was created by crowdsourcing well-formedness annotations for 25,100 queries from the Paralex corpus. Every query was annotated by five raters each with 1/0 rating of whether or not the query is well-formed.
Supported Tasks and Leaderboards
[More Information Needed]
Languages
English
Dataset Structure
Data Instances
{'rating': 0.2, 'content': 'The European Union includes how many ?'}
Data Fields
rating: afloatbetween 0-1sentence: query which you want to rate
Data Splits
| Train | Valid | Test | |
|---|---|---|---|
| Input Sentences | 17500 | 3750 | 3850 |
Dataset Creation
Curation Rationale
Understanding search queries is a hard problem as it involves dealing with “word salad” text ubiquitously issued by users. However, if a query resembles a well-formed question, a natural language processing pipeline is able to perform more accurate interpretation, thus reducing downstream compounding errors. Hence, identifying whether or not a query is well formed can enhance query understanding. This dataset introduce a new task of identifying a well-formed natural language question.
Source Data
Used the Paralex corpus (Fader et al., 2013) that contains pairs of noisy paraphrase questions. These questions were issued by users in WikiAnswers (a Question-Answer forum) and consist of both web-search query like constructs (“5 parts of chloroplast?”) and well-formed questions (“What is the punishment for grand theft?”).
Initial Data Collection and Normalization
Selected 25,100 queries from the unique list of queries extracted from the corpus such that no two queries in the selected set are paraphrases.
Who are the source language producers?
[More Information Needed]
Annotations
Annotation process
The queries are annotated into well-formed or non-wellformed questions if it satisfies the following:
- Query is grammatical.
- Query is an explicit question.
- Query does not contain spelling errors.
Who are the annotators?
Every query was labeled by five different crowdworkers with a binary label indicating whether a query is well-formed or not. And average of the ratings of the five annotators was reported, to get the probability of a query being well-formed.
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
[More Information Needed]
Licensing Information
Query-wellformedness dataset is licensed under CC BY-SA 4.0. Any third party content or data is provided “As Is” without any warranty, express or implied.
Citation Information
@InProceedings{FaruquiDas2018,
title = {{Identifying Well-formed Natural Language Questions}},
author = {Faruqui, Manaal and Das, Dipanjan},
booktitle = {Proc. of EMNLP},
year = {2018}
}
Contributions
Thanks to @vasudevgupta7 for adding this dataset.
- Downloads last month
- 102