变更可证监管：政策变更下学习制品的治理
Change-Provenant Supervision: Governing Learned Artifacts Under Policy Change
AI总结：
针对学习制品在政策变更下的治理，提出分离记录血缘与独立当前契约重新验证的机制，通过实验证明其能有效拒绝来源过期的包并恢复语义失效目标，实现受控治理。
AI中文摘要：
记录的依赖图无法证明其不包含遗漏的边。对于学习制品，图范围的失效因此无法在权限变更后单独证明接纳的合理性，尤其是当输出测试遗漏了来源过期的推导时。我们将记录的血缘（它提出影响范围和依赖解释）与对每个保留目标的独立当前契约重新验证（它提供相对于声明契约的可靠性）分开。我们在两个证据层面评估这一设计。四个Qwen3-14B LoRA包将1.028 GB的适配器字节绑定到训练和权限记录。在互斥的版本化监督上训练的过期和全新适配器在所有80个测试的权限中立输入上生成了相同的计划。然而，基于字节的重建和密封的当前契约账本拒绝了过期和朴素追加包，并接纳了全新和血缘选择性包。在一个具有六个角色分离修正的受控框架中，传递血缘恢复了所有40个语义无效的修正-目标对，同时提名了56个影响候选，相比之下，文档范围失效为120个。所有56个都有到修正条款的声明路径，尽管16个仍内容有效。每次修正省略一个承重边将语义召回率降至34/40，而独立重新验证拒绝了所有六个受影响的包。结果确立了一个受控治理机制的性质，而非企业失效率、遗忘结果、操作最优工作流或实测总成本降低。
英文摘要：
A recorded dependency graph cannot certify that it contains no omitted edge. For learned artifacts, graph-scoped invalidation therefore cannot by itself justify admission after authority changes, especially when output testing misses a provenance-stale derivation. We separate recorded lineage, which proposes impact scope and a dependency explanation, from independent current-contract revalidation of every retained target, which supplies soundness relative to the declared contract. We evaluate this design in two evidence layers. Four Qwen3-14B LoRA packages bind 1.028 GB of adapter bytes to training and authority records. Stale and full-fresh adapters trained on mutually exclusive versioned supervision emitted identical plans on all 80 tested authority-neutral inputs. Byte-derived reconstruction and a sealed current-contract ledger nevertheless refused the stale and naive-append packages and admitted full-fresh and lineage-selective packages. In a controlled harness with six role-separated amendments, transitive lineage recovered all 40 semantically invalidated amendment-target pairs while nominating 56 impact candidates, compared with 120 under document-wide invalidation. All 56 had a declared path to an amended clause, although 16 remained content-valid. Omitting one load-bearing edge per amendment reduced semantic recall to 34/40 while independent revalidation refused all six affected packages. The results establish properties of a controlled governance mechanism, not an enterprise invalidation rate, unlearning result, operationally optimal workflow, or measured total-cost reduction.
