Ad Hoc Information Retrieval On Trec Robust04

Metrics

MAP

P@20

nDCG@20

Results

Performance results of various models on this benchmark

Model Name	MAP	P@20	nDCG@20	Paper Title	Repository
monoT5-3B (zero-shot)	0.3876	0.5165	0.6091	Document Ranking with a Pretrained Sequence-to-Sequence Model
CEDR-KNRM	-	0.4667	0.5381	CEDR: Contextualized Embeddings for Document Ranking
FNRM-Rank_Embed	0.2811	-	-	Neural Ranking Models with Weak Supervision
DRMM	0.279	0.382	0.431	A Deep Relevance Matching Model for Ad-hoc Retrieval
FNRM-RankProb_Embed	0.2837	-	-	Neural Ranking Models with Weak Supervision
Vanilla BERT	-	0.4042	0.4541	CEDR: Contextualized Embeddings for Document Ranking
SNRM-PRF	0.2971	0.3948	0.4391	From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing	-
KNRM	0.2464	0.3510	0.3989	NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval
POSIT-DRMM-MV	0.271	0.389	0.464	Deep Relevance Ranking Using Enhanced Document-Query Interactions
PARADE(ELECTRA)	-	0.4604	0.5399	PARADE: Passage Representation Aggregation for Document Reranking
BERT-MaxP	-	-	0.469	Deeper Text Understanding for IR with Contextual Neural Language Modeling
Anserini BM25+RM3	0.302	0.4012	-	The Neural Hype and Comparisons Against Weak Baselines
SNRM	0.2856	0.3766	0.4310	From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing	-
BERT FT(Microblog)	0.3278	0.4287	-	Simple Applications of BERT for Ad Hoc Document Retrieval
PARADE(BERT)	-	0.4486	0.5252	PARADE: Passage Representation Aggregation for Document Reranking
NPRF-KNRM	0.2846	0.3926	0.4327	NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval
QL	0.2499	-	-	From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing	-
PACRR	0.258	0.374	0.445	Deep Relevance Ranking Using Enhanced Document-Query Interactions
BERT-SumP	-	-	0.467	Deeper Text Understanding for IR with Contextual Neural Language Modeling
NPRF-DRMM	0.2904	0.4064	0.4502	NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval

0 of 21 row(s) selected.