Reproduce “Beyond Homoscedasticity: Decoupled Uncertainty Optimization for Deep Imbalanced Regression” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Juncheng Zhou, Jiaxi Lu, Weijing Zeng, Zhong Li, Hao Qi, Jingsong Cui
DUO addresses deep imbalanced regression by predicting sample-specific uncertainty and using it to weight mean-prediction updates and guide contrastive feature pairing. Its decoupled losses and uncertainty-sensitive distribution overlap are evaluated on three scalar-regression benchmarks. On AgeDB-DIR few-shot examples, DUO’s bMAE is 11.590 versus LDS’s 12.055, and GM is 6.587 versus 7.577; few-shot MAE is not uniformly best.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.