DNR is an adversarial denoising neural reranker that extends score error minimization with three objectives to denoise retriever scores and align them with user feedback in two-stage recommender systems.
ISBN 9781450359405
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Denoising Neural Reranker for Recommender Systems
DNR is an adversarial denoising neural reranker that extends score error minimization with three objectives to denoise retriever scores and align them with user feedback in two-stage recommender systems.