AI Paper Atlas

Research map / Optimization / Optimizers

B2P-Corr: Batch-to-Population Gradient Estimators for Non-Decomposable Correlation Losses

Jingquan Yan, Yuwei Miao, Peiran Yu, Junzhou Huang

NeurIPS 2026 · Optimization › Optimizers · approach: Stochastic gradient MCMC · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes batch-to-population gradient estimators for training with non-decomposable correlation losses (PCC, CCC) under mini-batch stochastic optimization.

Problem

Correlation metrics such as PCC and CCC depend on dataset-level moments, so naive mini-batch training with local statistics gives biased, high-variance gradients.

Approach

Batch-to-population gradient estimators that correct mini-batch correlation gradients toward population-level statistics for non-decomposable correlation losses.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace naive mini-batch PCC with B2P-EMAKVQ · KoNViD-1k (batch size 2)PCC: 0.865 -> 0.895Improved
replace naive mini-batch PCC with B2P-EMAKVQ · KoNViD-1k (batch size 2)RMSE: 0.303 -> 0.283Improved
use B2P-EMA instead of the PLC fine-grained correlation lossKVQ · KoNViD-1k (batch size 2)PCC: 0.871 -> 0.895Improved
replace naive mini-batch PCC with B2P-UKVQ · KoNViD-1k (batch size 8)PCC: 0.881 -> 0.906Improved
replace naive PCC term with B2P-UEGN · 10xProteomic (batch size 8)PCC: 0.615 -> 0.660Improved
replace naive PCC term with B2P-UEGN · 10xProteomic (batch size 32, default)PCC: 0.629 -> 0.673Improved
replace naive CCC term with B2P-UVEATIC baseline · VEATIC valence (batch size 4)CCC: 0.437 -> 0.467Improved
replace naive CCC term with B2P-UVEATIC baseline · VEATIC valence (batch size 20, default)CCC: 0.606 -> 0.615Improved
use stop-gradient moment sketches instead of naive mini-batch PCC gradientsLinear predictor, PCC · Synthetic Exp.1 (B=4)gradient alignment: 0.16 -> 0.43Improved
add reservoir-based cross-batch U-statistic sketch (B2P-U vs B2P-EMA) (ablation)B2P-EMA · Synthetic Exp.1alignment std: 0.0275 -> 0.0183Improved
use B2P-U instead of B2P-EMA for mean gradient alignment (ablation)B2P-EMA · Synthetic Exp.1gradient alignmentNo clear effect
use B2P-EMA or B2P-U at batch size 2Linear predictor, PCC · Synthetic Exp.2 (B=2)DirectionRecovery, TestPCCImproved

Experimental setup

Baselines: naive mini-batch correlation loss

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy