HyperAIHyperAI

Command Palette

Search for a command to run...

Plateforme

Ad Hoc Information Retrieval On Trec Robust04

Métriques

MAP
P@20
nDCG@20

Résultats

Résultats de performance de divers modèles sur ce benchmark

Paper Title
monoT5-3B (zero-shot)0.38760.51650.6091Document Ranking with a Pretrained Sequence-to-Sequence Model
PARADE(ELECTRA)-0.46040.5399PARADE: Passage Representation Aggregation for Document Reranking
CEDR-KNRM-0.46670.5381CEDR: Contextualized Embeddings for Document Ranking
PARADE(BERT)-0.44860.5252PARADE: Passage Representation Aggregation for Document Reranking
BERT-MaxP--0.469Deeper Text Understanding for IR with Contextual Neural Language Modeling
BERT-SumP--0.467Deeper Text Understanding for IR with Contextual Neural Language Modeling
POSIT-DRMM-MV0.2710.3890.464Deep Relevance Ranking Using Enhanced Document-Query Interactions
Vanilla BERT-0.40420.4541CEDR: Contextualized Embeddings for Document Ranking
NPRF-DRMM0.29040.40640.4502NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval
PACRR0.2580.3740.445Deep Relevance Ranking Using Enhanced Document-Query Interactions
BERT-FirstP--0.444Deeper Text Understanding for IR with Contextual Neural Language Modeling
SNRM-PRF0.29710.39480.4391From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing
NPRF-KNRM0.28460.39260.4327NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval
DRMM0.2790.3820.431A Deep Relevance Matching Model for Ad-hoc Retrieval
SNRM0.28560.37660.4310From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing
KNRM0.24640.35100.3989NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval
FNRM-Rank_Embed0.2811--Neural Ranking Models with Weak Supervision
FNRM-RankProb_Embed0.2837--Neural Ranking Models with Weak Supervision
Anserini BM25+RM30.3020.4012-The Neural Hype and Comparisons Against Weak Baselines
BERT FT(Microblog)0.32780.4287-Simple Applications of BERT for Ad Hoc Document Retrieval
0 of 21 row(s) selected.