AI 中文总结
本研究提出 AccretionLink 框架，为属性推断的暴露控制攻击定义博弈并构建 e-过程，经多组画像实验验证其在降低负对数似然等方面的效果，且完成设备端相关摘要的身份验证。
AI 中文摘要
暴露控制使攻击者能够对真实的公开帖子进行排名，以在不改变内容的情况下强化私有属性推断。AccretionLink 为该攻击定义了保密性和完整性博弈，通过部分识别对有界选择几率进行建模，并构建了依赖感知的时间一致 e-过程。在 52 个预留的合成用户画像上，几率为四的选择在所有时间范围内均降低了聚合负对数似然；在 8 个帖子时，优势为 0.01595 纳特（95% 置信区间 [0.00890, 0.02336]），四分之三的目标效应通过 Holm 校正得以保留，且标签不可知的模型引导选择导致 109 个案例中出现 6 个高置信度的错误反转。在 142 个 PAN15 测试用户画像上，探索性选择产生了 0.01227 纳特的优势但无反转；单独的 TF-IDF 选择器针对未改变的 G5 目标保留了 0.01470 纳特的优势，而匹配的身份打乱未重现该优势。Pixel 10 使用无 fallback 的 Tensor G5 图对全部 1622 个预留帖子进行了编码；P-256 检查点对选定重放、实际模型、原生报告及操作摘要进行了身份验证，本地 KeyInfo 将签名密钥识别为 StrongBox 支持的。
英文摘要
Exposure control lets an adversary rank authentic public posts to strengthen private-attribute inference without altering content. AccretionLink defines confidentiality and integrity games for this attack, models bounded selection odds through partial identification, and constructs dependence-aware time-uniform e-processes. On 52 held-out synthetic profiles, odds-four selection reduced aggregate negative log likelihood at every horizon. At eight posts the advantage was 0.01595 nats (95% CI [0.00890, 0.02336]), three of four target effects survived Holm adjustment, and label-blind model-guided selection caused 6/109 high-confidence false reversals. On 142 PAN15 test profiles, exploratory selection produced a 0.01227-nat advantage but no reversal. A separate TF-IDF selector retained a 0.01470-nat advantage against the unchanged G5 target, while matched identity shuffling did not reproduce it. Pixel 10 encoded all 1,622 held-out posts once with a fallback-free Tensor G5 graph. A P-256 checkpoint authenticated the selected-replay, actual-model, native-report, and operation digests; local KeyInfo identified the signing key as StrongBox-backed.
Comments21 pages, 2 figures; includes proofs, reproducibility code, and Pixel 10/Tensor G5 verification artifacts as ancillary files
