Matches in SemOpenAlex for { <https://semopenalex.org/work/W4295682539> ?p ?o ?g. }
Showing items 1 to 75 of
75
with 100 items per page.
- W4295682539 abstract "Neural ranking models (NRMs) have become one of the most important techniques in information retrieval (IR). Due to the limitation of relevance labels, the training of NRMs heavily relies on negative sampling over unlabeled data. In general machine learning scenarios, it has shown that training with hard negatives (i.e., samples that are close to positives) could lead to better performance. Surprisingly, we find opposite results from our empirical studies in IR. When sampling top-ranked results (excluding the labeled positives) as negatives from a stronger retriever, the performance of the learned NRM becomes even worse. Based on our investigation, the superficial reason is that there are more false negatives (i.e., unlabeled positives) in the top-ranked results with a stronger retriever, which may hurt the training process; The root is the existence of pooling bias in the dataset constructing process, where annotators only judge and label very few samples selected by some basic retrievers. Therefore, in principle, we can formulate the false negative issue in training NRMs as learning from labeled datasets with pooling bias. To solve this problem, we propose a novel Coupled Estimation Technique (CET) that learns both a relevance model and a selection model simultaneously to correct the pooling bias for training NRMs. Empirical results on three retrieval benchmarks show that NRMs trained with our technique can achieve significant gains on ranking effectiveness against other baseline strategies." @default.
- W4295682539 created "2022-09-14" @default.
- W4295682539 creator A5006971161 @default.
- W4295682539 creator A5009898523 @default.
- W4295682539 creator A5029998682 @default.
- W4295682539 creator A5084642830 @default.
- W4295682539 creator A5088621320 @default.
- W4295682539 creator A5089655391 @default.
- W4295682539 date "2022-09-12" @default.
- W4295682539 modified "2023-10-18" @default.
- W4295682539 title "Hard Negatives or False Negatives: Correcting Pooling Bias in Training Neural Ranking Models" @default.
- W4295682539 doi "https://doi.org/10.48550/arxiv.2209.05072" @default.
- W4295682539 hasPublicationYear "2022" @default.
- W4295682539 type Work @default.
- W4295682539 citedByCount "0" @default.
- W4295682539 crossrefType "posted-content" @default.
- W4295682539 hasAuthorship W4295682539A5006971161 @default.
- W4295682539 hasAuthorship W4295682539A5009898523 @default.
- W4295682539 hasAuthorship W4295682539A5029998682 @default.
- W4295682539 hasAuthorship W4295682539A5084642830 @default.
- W4295682539 hasAuthorship W4295682539A5088621320 @default.
- W4295682539 hasAuthorship W4295682539A5089655391 @default.
- W4295682539 hasBestOaLocation W42956825391 @default.
- W4295682539 hasConcept C105795698 @default.
- W4295682539 hasConcept C112789634 @default.
- W4295682539 hasConcept C119857082 @default.
- W4295682539 hasConcept C124101348 @default.
- W4295682539 hasConcept C140779682 @default.
- W4295682539 hasConcept C153180895 @default.
- W4295682539 hasConcept C154945302 @default.
- W4295682539 hasConcept C158154518 @default.
- W4295682539 hasConcept C17744445 @default.
- W4295682539 hasConcept C189430467 @default.
- W4295682539 hasConcept C199539241 @default.
- W4295682539 hasConcept C33923547 @default.
- W4295682539 hasConcept C40423286 @default.
- W4295682539 hasConcept C41008148 @default.
- W4295682539 hasConcept C64869954 @default.
- W4295682539 hasConcept C70437156 @default.
- W4295682539 hasConcept C76155785 @default.
- W4295682539 hasConcept C94915269 @default.
- W4295682539 hasConceptScore W4295682539C105795698 @default.
- W4295682539 hasConceptScore W4295682539C112789634 @default.
- W4295682539 hasConceptScore W4295682539C119857082 @default.
- W4295682539 hasConceptScore W4295682539C124101348 @default.
- W4295682539 hasConceptScore W4295682539C140779682 @default.
- W4295682539 hasConceptScore W4295682539C153180895 @default.
- W4295682539 hasConceptScore W4295682539C154945302 @default.
- W4295682539 hasConceptScore W4295682539C158154518 @default.
- W4295682539 hasConceptScore W4295682539C17744445 @default.
- W4295682539 hasConceptScore W4295682539C189430467 @default.
- W4295682539 hasConceptScore W4295682539C199539241 @default.
- W4295682539 hasConceptScore W4295682539C33923547 @default.
- W4295682539 hasConceptScore W4295682539C40423286 @default.
- W4295682539 hasConceptScore W4295682539C41008148 @default.
- W4295682539 hasConceptScore W4295682539C64869954 @default.
- W4295682539 hasConceptScore W4295682539C70437156 @default.
- W4295682539 hasConceptScore W4295682539C76155785 @default.
- W4295682539 hasConceptScore W4295682539C94915269 @default.
- W4295682539 hasLocation W42956825391 @default.
- W4295682539 hasOpenAccess W4295682539 @default.
- W4295682539 hasPrimaryLocation W42956825391 @default.
- W4295682539 hasRelatedWork W2070589300 @default.
- W4295682539 hasRelatedWork W2291847203 @default.
- W4295682539 hasRelatedWork W2944724518 @default.
- W4295682539 hasRelatedWork W2951391129 @default.
- W4295682539 hasRelatedWork W2952858462 @default.
- W4295682539 hasRelatedWork W2999074369 @default.
- W4295682539 hasRelatedWork W3093454656 @default.
- W4295682539 hasRelatedWork W4225258897 @default.
- W4295682539 hasRelatedWork W4230430697 @default.
- W4295682539 hasRelatedWork W2472056907 @default.
- W4295682539 isParatext "false" @default.
- W4295682539 isRetracted "false" @default.
- W4295682539 workType "article" @default.