Decoupled Learning and Selection in Slate Recommendation for Privacy and Stability Under Noisy Scores

2026-09-24

AI Summary

This study addresses privacy leakage and noise-induced ranking instability in list recommendation by proposing a novel framework that decouples learning into stochastic scoring and deterministic selection. Methodologically, it constructs end-to-end privacy-propagating conditions alongside certifiable stability mechanisms, rigorously bounding unconditional and conditional stability guarantees through differential privacy and logarithmic margin analysis. Experimental results demonstrate that high anchor weights significantly suppress ranking fluctuations. Furthermore, empirical audits validate the effectiveness of the privacy contracts and the accuracy of the stability certificates. Overall, this work provides a theoretically rigorous yet practical solution for trustworthy recommendation systems.

Abstract

We formalize slate recommendation as a randomized score learner followed by deterministic selection. First, an appropriately scoped differential-privacy guarantee passes through selection and its audit trace by post-processing. End-to-end privacy holds only when selector inputs are public or independent, previous private outputs, or separately privacy-accounted; fixing raw state or candidate information instead yields only a conditional guarantee. Second, we derive a logged margin certificate: bounded score-induced objective movement below half the smallest greedy decision margin guarantees that the ordered slate is unchanged. Controlled fixed-margin tests show near-linear exponent scaling, with an empirical slope of $-0.220$ (95% CI $[-0.231,-0.210]$) against the independent-noise reference $-1/4$. Real-anchor experiments on OULAD, MovieLens-25M, and Amazon Musical Instruments show that greater anchor weight reduces score-noise-induced ranking churn. OULAD and EdNet certificate checks validate the implementation of the logged inequality, while closed-loop simulations show bounded target drift and setting-dependent downstream utility. The contribution is therefore a privacy-scope contract and a certifiable score-to-slate stability mechanism, not a universal utility claim.

Problem

• slate recommendation
• differential privacy
• score noise
• ranking stability
• margin certificate

Innovation

• Slate Recommendation
• Differential Privacy
• Decoupled Learning
• Margin Certificate
• Ranking Stability