OTSS: Output-Targeted Soft Segmentation for Contextual Decision-Weight Learning

TL;DR

OTSS introduces a novel soft segmentation model for contextual decision-weight learning, enabling personalized decision vectors with theoretical and empirical advantages over existing methods.

Contribution

The paper proposes OTSS, a soft segmentation approach that improves decision-weight learning by removing approximation limits and achieving faster, more accurate results.

Findings

OTSS attains the lowest mean regret in controlled benchmarks.

OTSS matches EM in coefficient recovery but is significantly faster.

OTSS performs well on real-world retail data, reducing regret.

Abstract

Many machine learning systems make constrained decisions by optimizing factorized objectives, but the context-specific objective is often treated as fixed. We study contextual decision-weight learning: from logged decisions and proxy outputs, learn an optimizer-facing weight vector w(x) over interpretable decision factors z(x,d), rather than a direct policy or generic predictive score. We propose OTSS, an output-targeted soft-segmentation model that deploys the personalized decision-ready weight vector. At the function-class level, the theory highlights a hard-versus-soft distinction. Hard partitions incur an approximation-estimation tradeoff under overlap, while a realizable fixed-K soft class removes the hard-partition approximation floor and attains a parametric rate. We evaluate OTSS in controlled benchmarks with finite evaluation libraries, where the true weight vector and downstream regret can be computed exactly. In the representative overlap setting, OTSS attains the lowest mean regret among the comparators, including EM mixture regression, the strongest soft-mixture baseline in our comparison; it matches EM on coefficient recovery while running about two orders of magnitude faster. In a matched K=5 benchmark, OTSS remains competitive under hard-routed truth and improves as heterogeneity becomes softer and sample size grows. On a fixed Complete Journey retail anchor with real household covariates and action geometry, OTSS again achieves the lowest mean-regret point estimate.