Periodically refreshing hard negatives with the in-training retriever improves unsupervised domain adaptation of dense retrievers over the static GPL baseline.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Remining Hard Negatives for Generative Pseudo Labeled Domain Adaptation
Periodically refreshing hard negatives with the in-training retriever improves unsupervised domain adaptation of dense retrievers over the static GPL baseline.