quac

  • תיאור :

תשובות לשאלות בהקשר הוא מערך נתונים ליצירת מודלים, הבנה והשתתפות בדיאלוג חיפוש מידע. מופעי נתונים מורכבים מדיאלוג אינטראקטיבי בין שני עובדי המונים: (1) תלמיד שמציג רצף של שאלות בצורה חופשית כדי ללמוד כמה שיותר על טקסט מוסתר בוויקיפדיה, ו-(2) מורה שעונה על השאלות על ידי אספקת קטעים קצרים (מתפרש) מהטקסט. QuAC מציגה אתגרים שלא נמצאים במערכי נתונים קיימים של הבנת המכונה: השאלות שלו הן לעתים קרובות יותר פתוחות, בלתי ניתנות לתשובה, או רק בעלות משמעות בהקשר הדו-שיח.

לְפַצֵל דוגמאות
'train' 83,568
'validation' 7,354
  • מבנה תכונה :
FeaturesDict({
    'answers': Sequence({
        'answer_start': int32,
        'text': Text(shape=(), dtype=string),
    }),
    'background': Text(shape=(), dtype=string),
    'context': Text(shape=(), dtype=string),
    'followup': Text(shape=(), dtype=string),
    'orig_answer': FeaturesDict({
        'answer_start': int32,
        'text': Text(shape=(), dtype=string),
    }),
    'question': Text(shape=(), dtype=string),
    'section_title': Text(shape=(), dtype=string),
    'title': Text(shape=(), dtype=string),
    'yesno': Text(shape=(), dtype=string),
})
  • תיעוד תכונה :
תכונה מעמד צוּרָה Dtype תיאור
FeaturesDict
תשובות סדר פעולות
answers/answer_start מוֹתֵחַ int32
תשובות/טקסט טֶקסט חוּט
רקע כללי טֶקסט חוּט
הֶקשֵׁר טֶקסט חוּט
תתעדכן טֶקסט חוּט
מקור_תשובה FeaturesDict
orig_answer/answer_start מוֹתֵחַ int32
orig_answer/text טֶקסט חוּט
שְׁאֵלָה טֶקסט חוּט
section_title טֶקסט חוּט
כותרת טֶקסט חוּט
כן לא טֶקסט חוּט
  • ציטוט :
@article{choi2018quac,
  title={Quac: Question answering in context},
  author={Choi, Eunsol and He, He and Iyyer, Mohit and Yatskar, Mark and Yih, Wen-tau and Choi, Yejin and Liang, Percy and Zettlemoyer, Luke},
  journal={arXiv preprint arXiv:1808.07036},
  year={2018}
}