Dataset Viewer
Auto-converted to Parquet Duplicate
passage_answer_candidates
dict
question_text
stringlengths
1
226
document_title
stringlengths
1
116
language
stringclasses
11 values
annotations
dict
document_plaintext
stringlengths
11
739k
document_url
stringlengths
31
792
{ "plaintext_start_byte": [ 1, 660, 844, 1196, 1354, 2126, 3608, 4198, 4509, 4741, 6531, 7666, 8000, 8562, 9249, 9616, 10774, 11127, 11980, 12747, 13325, 13503, 15400, 17459 ], "plaintext_end_byte": [ 659, 843,...
berapakah jenis ras yang ada didunia?
Ras manusia
indonesian
{ "passage_answer_candidate_index": [ -1 ], "minimal_answers_start_byte": [ -1 ], "minimal_answers_end_byte": [ -1 ], "yes_no_answer": [ "NONE" ] }
transl. Ras (dari bahasa Prancis race, yang sendirinya dari bahasa Latin radix, "akar") adalah suatu sistem klasifikasi yang digunakan untuk mengkategorikan manusia dalam populasi atau kelompok besar dan berbeda melalui ciri fenotipe, asal usul geografis, tampang jasmani dan kesukuan yang terwarisi. Di awal abad ke-2...
https://id.wikipedia.org/wiki/Ras%20manusia
{ "plaintext_start_byte": [ 1, 271, 995, 1763, 2415, 2962, 3546, 4259, 5100, 5654, 6040, 6364, 6741, 7991, 9031, 9276, 9780, 10251, 10950, 11043, 11583, 11897, 12084, 12447, 13148, 13560, 14430, 15112, ...
2018年アメリカで一番治安の悪い州はどこ
デトロイト
japanese
{ "passage_answer_candidate_index": [ -1 ], "minimal_answers_start_byte": [ -1 ], "minimal_answers_end_byte": [ -1 ], "yes_no_answer": [ "NONE" ] }
デトロイト( /dɨˈtrɔɪt/)は、アメリカ合衆国ミシガン州南東部にある都市。南北をエリー湖とヒューロン湖に挟まれており、東はカナダのウィンザー市に接する。アメリカ中西部有数の世界都市。 人口は、2000年国勢調査では951,270人、2012年では701,475人と減り続けている。デトロイトの都市圏(大都市統計地域:MSA)の人口は4,452,559人であり、全米第9位の規模。フリントなどを含めた広域都市圏(合同統計地域:CSA)の人口は5,357,538人(いずれも2000年国勢調査)で、やはり全米9位の規模である。また、失業率、貧困率が高く、犯罪都市としても知られていた。主要産業は自動車産業であり、「自動車の街」とも...
https://ja.wikipedia.org/wiki/%E3%83%87%E3%83%88%E3%83%AD%E3%82%A4%E3%83%88
{ "plaintext_start_byte": [ 0, 208, 542, 891, 1006, 1348, 1910, 2100, 2562, 2863, 3067, 3237, 3525, 3869, 4118, 4320, 4831, 5078, 5235, 5558, 5678, 5747, 5841 ], "plaintext_end_byte": [ 207, 541, 890, 1005,...
Je,Ngamia anaweza kaa bila maji kwa muda gani?
Kuku
swahili
{ "passage_answer_candidate_index": [ -1 ], "minimal_answers_start_byte": [ -1 ], "minimal_answers_end_byte": [ -1 ], "yes_no_answer": [ "NONE" ] }
Kuku (Gallus gallus domesticus) ni ndege anayefugwa na binadamu nyumbani tangu miaka 8,000 hivi. Kwa sababu hiyo ni miongoni mwa wanyama waliosambaa zaidi duniani na ndio ndege walio wengi zaidi duniani.[1] Watu hutumia nyama yake na mayai kama chakula. Mwanzoni kuku walifungwa kwa ajili ya mchezo wa mapigano ya kuku ...
https://sw.wikipedia.org/wiki/Kuku
{ "plaintext_start_byte": [ 5, 401, 1185, 2103, 3527, 5959, 6460, 8338, 11805, 13221, 15645, 16550, 18690, 19961, 22283, 24460, 26218, 26527, 29887, 30212, 30956, 31299, 32383, 35082, 37489, 38110, 39694, 4...
কম্পিউটার বিজ্ঞানের মোট কয়টি শাখা রয়েছে ?
বিজ্ঞান
bengali
{ "passage_answer_candidate_index": [ -1 ], "minimal_answers_start_byte": [ -1 ], "minimal_answers_end_byte": [ -1 ], "yes_no_answer": [ "NONE" ] }
ভৌত বিশ্বের যা কিছু পর্যবেক্ষণযোগ্য, পরীক্ষণযোগ্য ও যাচাইযোগ্য, তার সুশৃঙ্খল, নিয়মতান্ত্রিক গবেষণা ও সেই গবেষণালব্ধ জ্ঞানভাণ্ডারের নাম বিজ্ঞান। ল্যাটিন শব্দ সায়েনটিয়া (scientia) থেকে ইংরেজি সায়েন্স শব্দটি এসেছে, যার অর্থ হচ্ছে জ্ঞান। বাংলা ভাষায় বিজ্ঞান শব্দটির অর্থ বিশেষ জ্ঞান।[2] ধারাবাহিক পর্যবেক্ষণ ও গ...
https://bn.wikipedia.org/wiki/%E0%A6%AC%E0%A6%BF%E0%A6%9C%E0%A7%8D%E0%A6%9E%E0%A6%BE%E0%A6%A8
{"plaintext_start_byte":[0,395,716,1320,3958,4136,4482,4782,6170,6569,7148,7476,8107,8967,9731,9941,(...TRUNCATED)
మెదక్ నగర విస్తీర్ణం ఎంత?
మెదక్ జిల్లా
telugu
{"passage_answer_candidate_index":[-1],"minimal_answers_start_byte":[-1],"minimal_answers_end_byte":(...TRUNCATED)
"మెదక్ జిల్లా తెలంగాణ రాష్ట్రంలోని 31 (...TRUNCATED)
https://te.wikipedia.org/wiki/%E0%B0%AE%E0%B1%86%E0%B0%A6%E0%B0%95%E0%B1%8D%20%E0%B0%9C%E0%B0%BF%E0%B0%B2%E0%B1%8D%E0%B0%B2%E0%B0%BE
{"plaintext_start_byte":[0,890,1247,1764,2218,3193,4384,4911,5372,5672,7072,7555,8034,8565,8928,9355(...TRUNCATED)
嗜好する音楽のジャンルと学業パフォーマンスに関連性はある?
音楽選好の心理学
japanese
{"passage_answer_candidate_index":[-1],"minimal_answers_start_byte":[-1],"minimal_answers_end_byte":(...TRUNCATED)
"音楽選好の心理学(おんがくせんこうのしんりがく)は、人それぞれの音(...TRUNCATED)
https://ja.wikipedia.org/wiki/%E9%9F%B3%E6%A5%BD%E9%81%B8%E5%A5%BD%E3%81%AE%E5%BF%83%E7%90%86%E5%AD%A6
{"plaintext_start_byte":[1,206,1144,2510,3680,3866,4725,6388,7379,8536,8989,9119,9564,9756,9825],"pl(...TRUNCATED)
"పనుకురాతిపాలెం గ్రామానికి దక్షిణాన (...TRUNCATED)
పనుకురాతిపాలెం
telugu
{"passage_answer_candidate_index":[-1],"minimal_answers_start_byte":[-1],"minimal_answers_end_byte":(...TRUNCATED)
"\nపనుకురాతిపాలెం, తూర్పు గోదావరి జిల్(...TRUNCATED)
https://te.wikipedia.org/wiki/%E0%B0%AA%E0%B0%A8%E0%B1%81%E0%B0%95%E0%B1%81%E0%B0%B0%E0%B0%BE%E0%B0%A4%E0%B0%BF%E0%B0%AA%E0%B0%BE%E0%B0%B2%E0%B1%86%E0%B0%82
{"plaintext_start_byte":[1,415,1004,1307,1962,2613,3328,3395,4168,4224,4951,5732,6551,7214,8340,8746(...TRUNCATED)
Dari manakah asal Dr. H. Shofwan Karim Elha?
Universitas Islam Negeri Syarif Hidayatullah Jakarta
indonesian
{"passage_answer_candidate_index":[-1],"minimal_answers_start_byte":[-1],"minimal_answers_end_byte":(...TRUNCATED)
"\nUniversitas Islam Negeri Syarif Hidayatullah Jakarta atau UIN Jakarta (sebelumnya: IAIN Syarif Hi(...TRUNCATED)
https://id.wikipedia.org/wiki/Universitas%20Islam%20Negeri%20Syarif%20Hidayatullah%20Jakarta
{"plaintext_start_byte":[1,163,1213,1444,1583,2165,2706,3304,4292,4722,5517,6079,6409,6868,7207,8076(...TRUNCATED)
Milloin Charles Fort syntyi?
Charles Fort
finnish
{"passage_answer_candidate_index":[0],"minimal_answers_start_byte":[19],"minimal_answers_end_byte":[(...TRUNCATED)
"\nCharles Hoy Fort (6. elokuuta (joidenkin lähteiden mukaan 9.) 1874 – 3. toukokuuta 1932) oli y(...TRUNCATED)
https://fi.wikipedia.org/wiki/Charles%20Fort
{"plaintext_start_byte":[4,286,1134,2256,3092,3279,3768,4260,4547,4897,4930,5067,5500,5817,6017,6258(...TRUNCATED)
siapakah pastur/ketua Ibadah pertama GBI KA?
GBI Keluarga Allah
indonesian
{"passage_answer_candidate_index":[-1],"minimal_answers_start_byte":[-1],"minimal_answers_end_byte":(...TRUNCATED)
"\n\n\n\nGBI Keluarga Allah atau singkatnya GBI KA adalah sebuah gereja kristen di Indonesia yang be(...TRUNCATED)
https://id.wikipedia.org/wiki/GBI%20Keluarga%20Allah
End of preview. Expand in Data Studio

google-research-datasets/tydiqa(複製)

データの内容

類型的に多様な 11 言語の、記事の本文を読んで答える質問応答のデータ。1 行が 1 件の質問で、question_text に質問文、document_plaintext に元の文書、annotations に答えの位置と Yes/No の回答が入る。質問は答えを知らない人が書き、翻訳を使わず各言語で直接集められた。構成は primary_task と secondary_task の 2 つ。

データの概要

項目 値
行数 240,544(2 構成・4 分割の合計)
列数 5〜7(構成による)
容量 2.94 GB(ファイル 17 本)
ファイルの形式 parquet 15
言語 ar, bn, en, fi, id, ja, ko, ru, sw, te, th
公開 public

構成と分割

構成(config) 分割(split)ごとの行数 列数
primary_task train 166,916、validation 18,670 7
secondary_task train 49,881、validation 5,077 5

列(構成 primary_task・分割 train。構成ごとに列が違う)

列 型 先頭の行の値
passage_answer_candidates list<struct{plaintext_start_byte: int32, plaintext_end_byte: int32}> {plaintext_start_byte, plaintext_end_byte}
question_text string berapakah jenis ras yang ada didunia?
document_title string Ras manusia
language string indonesian
annotations list<struct{passage_answer_candidate_index: int32, minimal_answers_start_byte: int32, minimal_answers_end_byte: int32, yes_no_answer: string}> {passage_answer_candidate_index, minimal_answers_start_byte, minimal_answers_end_byte, yes_no_answer}
document_plaintext string transl. Ras (dari bahasa Prancis race, yang sen…
document_url string https://id.wikipedia.org/wiki/Ras%20manusia

行数と列は HF のデータセットビューアの集計(元のデータセットの現在の版で数えたもの)。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/tydiqa をリビジョン da78f23f9119363459acbaf46bf89426ff26c259 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

論文

  • 元のカードに論文へのリンクは見つからなかった

アノテーションの方法

作成の根拠

More Information Needed

最初のデータ収集と正規化

More Information Needed

元の言語データの作成者は誰か?

More Information Needed

アノテーションの過程

More Information Needed

データセットの作成者

More Information Needed

微調整に使うときの注意

  • 調査メモの台帳での区分: A(許諾型のライセンスだけで、疑義の注記が無い)。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
multilingual_cjk.md 1. 大規模多言語データセット(XNLI, PAWS-X, MASSIVE, SIB-200, Belebele ほか) (L52) apache-2.0 確認済 確認印なし(各ファイルの凡例により確認済扱い) / 原文注記: TyDi の primary_task に yes/no 回答欄(yes_no_answer)がある。HF card で確認できた値は NONE のみで、YES/NO の件数は未確認 → noul 候補だが件数要確認

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。


"tydiqa" のデータセットカード

目次

  • データセットの説明
    • データセットの概要
    • 対象タスクとリーダーボード
    • 言語
  • データセットの構造
    • データインスタンス
    • データフィールド
    • データ分割
  • データセットの作成
    • 作成の根拠
    • 元データ
    • アノテーション
    • 個人情報と機微な情報
  • データ利用上の留意点
    • データセットの社会的影響
    • バイアスの検討
    • 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

TyDi QA は、類型的に多様な 11 言語にわたる、204K 組の質問と回答のペアを含む質問応答データセットである。TyDi QA の言語は、類型論(各言語が表す言語的な特徴の集合)の点で多様である。そのため、このセットでうまく動作するモデルは、世界の多数の言語へ一般化することが期待できる。英語だけのコーパスには見られない言語現象も含む。現実的な情報探索のタスクを用意し、プライミング効果を避けるため、質問は答えを知りたいがまだ答えを知らない人が書いている(SQuAD やその派生データセットとは異なる)。また、データは翻訳を使わず、各言語で直接収集されている(MLQA や XQuAD とは異なる)。

対象タスクとリーダーボード

More Information Needed

言語

More Information Needed

データセットの構造

データインスタンス

primary_task

  • ダウンロードしたデータセットファイルのサイズ: 1.95 GB
  • 生成されたデータセットのサイズ: 6.04 GB
  • 使用ディスク容量の合計: 7.99 GB

'validation' の例は次のとおり。

This example was too long and was cropped:

{
    "annotations": {
        "minimal_answers_end_byte": [-1, -1, -1],
        "minimal_answers_start_byte": [-1, -1, -1],
        "passage_answer_candidate_index": [-1, -1, -1],
        "yes_no_answer": ["NONE", "NONE", "NONE"]
    },
    "document_plaintext": "\"\\nรองศาสตราจารย์[1] หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร  (22 กันยายน 2495 -) ผู้ว่าราชการกรุงเทพมหานครคนที่ 15 อดีตรองหัวหน้าพรรคปร...",
    "document_title": "หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร",
    "document_url": "\"https://th.wikipedia.org/wiki/%E0%B8%AB%E0%B8%A1%E0%B9%88%E0%B8%AD%E0%B8%A1%E0%B8%A3%E0%B8%B2%E0%B8%8A%E0%B8%A7%E0%B8%87%E0%B8%...",
    "language": "thai",
    "passage_answer_candidates": "{\"plaintext_end_byte\": [494, 1779, 2931, 3904, 4506, 5588, 6383, 7122, 8224, 9375, 10473, 12563, 15134, 17765, 19863, 21902, 229...",
    "question_text": "\"หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร เรียนจบจากที่ไหน ?\"..."
}

secondary_task

  • ダウンロードしたデータセットファイルのサイズ: 1.95 GB
  • 生成されたデータセットのサイズ: 58.03 MB
  • 使用ディスク容量の合計: 2.01 GB

'validation' の例は次のとおり。

This example was too long and was cropped:

{
    "answers": {
        "answer_start": [394],
        "text": ["بطولتين"]
    },
    "context": "\"أقيمت البطولة 21 مرة، شارك في النهائيات 78 دولة، وعدد الفرق التي فازت بالبطولة حتى الآن 8 فرق، ويعد المنتخب البرازيلي الأكثر تت...",
    "id": "arabic-2387335860751143628-1",
    "question": "\"كم عدد مرات فوز الأوروغواي ببطولة كاس العالم لكرو القدم؟\"...",
    "title": "قائمة نهائيات كأس العالم"
}

データフィールド

すべての分割でデータフィールドは同じである。

primary_task

  • passage_answer_candidates: 次を含む辞書型の特徴量:
    • plaintext_start_byte: int32 型の特徴量。
    • plaintext_end_byte: int32 型の特徴量。
  • question_text: string 型の特徴量。
  • document_title: string 型の特徴量。
  • language: string 型の特徴量。
  • annotations: 次を含む辞書型の特徴量:
    • passage_answer_candidate_index: int32 型の特徴量。
    • minimal_answers_start_byte: int32 型の特徴量。
    • minimal_answers_end_byte: int32 型の特徴量。
    • yes_no_answer: string 型の特徴量。
  • document_plaintext: string 型の特徴量。
  • document_url: string 型の特徴量。

secondary_task

  • id: string 型の特徴量。
  • title: string 型の特徴量。
  • context: string 型の特徴量。
  • question: string 型の特徴量。
  • answers: 次を含む辞書型の特徴量:
    • text: string 型の特徴量。
    • answer_start: int32 型の特徴量。

データ分割

name train validation
primary_task 166916 18670
secondary_task 49881 5077

データセットの作成

作成の根拠

More Information Needed

元データ

最初のデータ収集と正規化

More Information Needed

元の言語データの作成者は誰か?

More Information Needed

アノテーション

アノテーションの過程

More Information Needed

アノテーターは誰か?

More Information Needed

個人情報と機微な情報

More Information Needed

データ利用上の留意点

データセットの社会的影響

More Information Needed

バイアスの検討

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

More Information Needed

引用情報

@article{tydiqa,
title   = {TyDi QA: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages},
author  = {Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}
year    = {2020},
journal = {Transactions of the Association for Computational Linguistics}
}

貢献者

このデータセットの追加に協力してくれた @thomwolf、@albertvillanova、@lewtun、@patrickvonplaten に感謝する。

原文(英語)

Dataset Card for "tydiqa"

Dataset Summary

TyDi QA is a question answering dataset covering 11 typologically diverse languages with 204K question-answer pairs. The languages of TyDi QA are diverse with regard to their typology -- the set of linguistic features that each language expresses -- such that we expect models performing well on this set to generalize across a large number of the languages in the world. It contains language phenomena that would not be found in English-only corpora. To provide a realistic information-seeking task and avoid priming effects, questions are written by people who want to know the answer, but don’t know the answer yet, (unlike SQuAD and its descendents) and the data is collected directly in each language without the use of translation (unlike MLQA and XQuAD).

Supported Tasks and Leaderboards

More Information Needed

Languages

More Information Needed

Dataset Structure

Data Instances

primary_task

  • Size of downloaded dataset files: 1.95 GB
  • Size of the generated dataset: 6.04 GB
  • Total amount of disk used: 7.99 GB

An example of 'validation' looks as follows.

This example was too long and was cropped:

{
    "annotations": {
        "minimal_answers_end_byte": [-1, -1, -1],
        "minimal_answers_start_byte": [-1, -1, -1],
        "passage_answer_candidate_index": [-1, -1, -1],
        "yes_no_answer": ["NONE", "NONE", "NONE"]
    },
    "document_plaintext": "\"\\nรองศาสตราจารย์[1] หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร  (22 กันยายน 2495 -) ผู้ว่าราชการกรุงเทพมหานครคนที่ 15 อดีตรองหัวหน้าพรรคปร...",
    "document_title": "หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร",
    "document_url": "\"https://th.wikipedia.org/wiki/%E0%B8%AB%E0%B8%A1%E0%B9%88%E0%B8%AD%E0%B8%A1%E0%B8%A3%E0%B8%B2%E0%B8%8A%E0%B8%A7%E0%B8%87%E0%B8%...",
    "language": "thai",
    "passage_answer_candidates": "{\"plaintext_end_byte\": [494, 1779, 2931, 3904, 4506, 5588, 6383, 7122, 8224, 9375, 10473, 12563, 15134, 17765, 19863, 21902, 229...",
    "question_text": "\"หม่อมราชวงศ์สุขุมพันธุ์ บริพัตร เรียนจบจากที่ไหน ?\"..."
}

secondary_task

  • Size of downloaded dataset files: 1.95 GB
  • Size of the generated dataset: 58.03 MB
  • Total amount of disk used: 2.01 GB

An example of 'validation' looks as follows.

This example was too long and was cropped:

{
    "answers": {
        "answer_start": [394],
        "text": ["بطولتين"]
    },
    "context": "\"أقيمت البطولة 21 مرة، شارك في النهائيات 78 دولة، وعدد الفرق التي فازت بالبطولة حتى الآن 8 فرق، ويعد المنتخب البرازيلي الأكثر تت...",
    "id": "arabic-2387335860751143628-1",
    "question": "\"كم عدد مرات فوز الأوروغواي ببطولة كاس العالم لكرو القدم؟\"...",
    "title": "قائمة نهائيات كأس العالم"
}

Data Fields

The data fields are the same among all splits.

primary_task

  • passage_answer_candidates: a dictionary feature containing:
    • plaintext_start_byte: a int32 feature.
    • plaintext_end_byte: a int32 feature.
  • question_text: a string feature.
  • document_title: a string feature.
  • language: a string feature.
  • annotations: a dictionary feature containing:
    • passage_answer_candidate_index: a int32 feature.
    • minimal_answers_start_byte: a int32 feature.
    • minimal_answers_end_byte: a int32 feature.
    • yes_no_answer: a string feature.
  • document_plaintext: a string feature.
  • document_url: a string feature.

secondary_task

  • id: a string feature.
  • title: a string feature.
  • context: a string feature.
  • question: a string feature.
  • answers: a dictionary feature containing:
    • text: a string feature.
    • answer_start: a int32 feature.

Data Splits

name train validation
primary_task 166916 18670
secondary_task 49881 5077

Dataset Creation

Curation Rationale

More Information Needed

Source Data

Initial Data Collection and Normalization

More Information Needed

Who are the source language producers?

More Information Needed

Annotations

Annotation process

More Information Needed

Who are the annotators?

More Information Needed

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

More Information Needed

Citation Information

@article{tydiqa,
title   = {TyDi QA: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages},
author  = {Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}
year    = {2020},
journal = {Transactions of the Association for Computational Linguistics}
}

Contributions

Thanks to @thomwolf, @albertvillanova, @lewtun, @patrickvonplaten for adding this dataset.

Downloads last month
69