paper-with-me

Question Answering 벤치마크

Question Answering on ConditionalQA

3개 결과 · ⬇ CSV · JSON

Conditional (answers) 낮을수록 좋음

39.4 40.85 42.3 43.75 45.2 2020-04 2026-09 ETC-Pipeline — 39.4 (2020-04-17) FiD — 45.2 (2020-07-02) DocHopper — 42.0 (2021-06-01) ETC-Pipeline — 39.4 (2020-04-17)
RankModel Conditional (answers)Conditional (w/ conditions)Overall (answers)Overall (w/ conditions) PaperCodeYear
1 FiD 45.2 / 49.74.7 / 5.844.4 / 50.835.0 / 40.6 Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering jhyuklee/DensePhrases · princeton-nlp/DensePhrases · facebookresearch/FiD · +5 2020
2 DocHopper 42.0 / 46.43.1 / 3.840.6 / 45.231.9 / 36.0 Iterative Hierarchical Attention for Answering Complex Questions over Long Documents 2021
3 ETC-Pipeline 39.4 / 41.82.5 / 3.435.6 / 39.826.9 / 30.8 ETC: Encoding Long and Structured Inputs in Transformers google-research/google-research · google-research/longt5 2020
1–3 / 3 페이지당 10 20 50 100