Papers

arXiv 论文提出 Serialization Tax:逐项打分导致群体决策失准并给出量化框架

Serialization Tax in Shared-Latent Exchangeable Decisions

Why we picked it

这篇 arXiv 论文给一个容易被忽略的坑起了名字:逐条概率都对,批量决策照样错。作者用凸序和 MovieLens 实验把误差量化了,还给出 TaxScore 做路由的落地方法。做推荐系统、批量审核、群体级决策的工程师应该看看。

论文《Serialization Tax in Shared-Latent Exchangeable Decisions》研究一个反直觉现象:逐项概率准确,群体层面决策仍可能出错。当多个条目共享一个不确定的隐含状态(如批量审核、面板展示、推迟决策)时,独立边缘化虽保留均值,却把依赖性从聚合后验中抹掉,论文将这一误差命名为 serialization tax。

Body

机制上,作者用经典凸序比较证明:共享同一成功率的二元条目,其精确计数分布比任何均值相同的独立补全都更分散,并进一步量化方差与全变差差距、在边际条件下推导阈值分离,把后验近似误差与决策 regret 关联起来,还覆盖了非二元结果、异质条目与后验充分分辨的情形。

实验部分,TaxScore 用于识别靠近决策边界的大块未解不确定性条目,路由到更强的后验头。在预设的、无隐藏目标的 MovieLens 1M/10M 审计中,现有 shared-latent heads 把高分 NLL/效用从 0.673/0.696 提升到 0.629/0.723、从 0.670/0.691 提升到 0.639/0.715,路由带来 0.006–0.012 的全群体效用增益。

Original · arXiv cs.AI

Serialization Tax in Shared-Latent Exchangeable Decisions

Accurate item probabilities can produce inaccurate group-level decisions. A system may score each item separately, yet act on whether enough hidden items succeed: show a slate, defer a panel decision, or inspect a batch. When items share an uncertain latent state, independent marginalization preserves means but removes dependence from the aggregate posterior. We call this error the serialization tax. For binary items sharing one success rate, classical convex-order comparisons give a more dispersed exact count law than any independent completion with the same total mean. We quantify variance and total-variation gaps, derive threshold separation under margin conditions, and connect posterior approximation to decision regret. Separate results cover non-binary outcomes, heterogeneous items, and sufficient posterior resolution. TaxScore identifies large hidden blocks with unresolved uncertainty near a decision boundary for routing to richer posterior heads. On pre-specified, hidden-target-free MovieLens 1M/10M audits, existing shared-latent heads improve high-score negative log-likelihood (NLL)/utility from 0.673/0.696 to 0.629/0.723 and from 0.670/0.691 to 0.639/0.715. Routing gives 0.006--0.012 full-population utility gains, with matched controls identifying their sources. Our code is available at https://github.com/SiminChang/Serialization-Tax-in-Shared-Latent-Exchangeable-Decisions.

Industry Coordinates & Source Verification

This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include MovieLens, recommendation systems, TaxScore.

Source attribution: curated from arXiv cs.AI (@Siming Zhang, Zhehui Shen, Shijie Chen, Xinle Gu, Yansen Yu, Hang Yu).