DADF performs second-stage multiplicative residual correction on watch-time predictors using dynamic distribution-aware transformation, debias-factor-aware module with video duration, and multi-label-aware module to mitigate local calibration bias in long-tailed targets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
DADF: A Distribution-Aware Debiasing Framework for Watch-Time Regression in Recommender Systems
DADF performs second-stage multiplicative residual correction on watch-time predictors using dynamic distribution-aware transformation, debias-factor-aware module with video duration, and multi-label-aware module to mitigate local calibration bias in long-tailed targets.