📄 中文摘要
研究提出了一种决策值映射的概念，用于记录不同数据表征对离散结果的影响。不同的表征可能会导致相同数据产生不同的离散结果，其中某些表征能够保留结果，而另一些则可能完全改变结果。决策值映射将每个声明的表征家族成员与其产生的离散结果关联起来。为此，构建了DecisionDB基础设施，能够记录、重放和审计这些关系，利用从内容和以一次写入形式存储的工件中计算出的标识符。确定性重放能够从存储的工件中精确恢复每个记录的决策标识符，确保所有三个标识字段与其匹配。
决策值映射与表征依赖性
出处: On Decision-Valued Maps and Representational Dependence
发布: 2026年2月13日
📄 English Summary
On Decision-Valued Maps and Representational Dependence
The study introduces the concept of decision-valued maps, which are used to record the impact of different representations of data on discrete outcomes. Different representations can lead to varying discrete results from the same data, with some preserving the outcome while others may alter it entirely. Decision-valued maps associate each member of a declared representation family with the discrete result it produces. To facilitate this, an infrastructure called DecisionDB is established, which logs, replays, and audits these relationships using identifiers computed from content and artifacts stored in a write-once format. Deterministic replay accurately recovers each recorded decision identifier from stored artifacts, ensuring that all three identifying fields match.