Ad Hoc Information Retrieval On Trec Robust04

MAP

P@20

nDCG@20

평가 결과

이 벤치마크에서 각 모델의 성능 결과

비교 표

모델 이름	MAP	P@20	nDCG@20
document-ranking-with-a-pretrained-sequence	0.3876	0.5165	0.6091
190407094	-	0.4667	0.5381
neural-ranking-models-with-weak-supervision	0.2811	-	-
a-deep-relevance-matching-model-for-ad-hoc	0.279	0.382	0.431
neural-ranking-models-with-weak-supervision	0.2837	-	-
190407094	-	0.4042	0.4541
from-neural-re-ranking-to-neural-ranking	0.2971	0.3948	0.4391
nprf-a-neural-pseudo-relevance-feedback	0.2464	0.3510	0.3989
deep-relevance-ranking-using-enhanced	0.271	0.389	0.464
parade-passage-representation-aggregation-for	-	0.4604	0.5399
deeper-text-understanding-for-ir-with	-	-	0.469
the-neural-hype-and-comparisons-against-weak	0.302	0.4012	-
from-neural-re-ranking-to-neural-ranking	0.2856	0.3766	0.4310
simple-applications-of-bert-for-ad-hoc	0.3278	0.4287	-
parade-passage-representation-aggregation-for	-	0.4486	0.5252
nprf-a-neural-pseudo-relevance-feedback	0.2846	0.3926	0.4327
from-neural-re-ranking-to-neural-ranking	0.2499	-	-
deep-relevance-ranking-using-enhanced	0.258	0.374	0.445
deeper-text-understanding-for-ir-with	-	-	0.467
nprf-a-neural-pseudo-relevance-feedback	0.2904	0.4064	0.4502
deeper-text-understanding-for-ir-with	-	-	0.444