发表机构
OPPO Health Lab; Guangdong OPPO Mobile Telecommunications Corp., Ltd.; Peking University; Xidian University; University of the Chinese Academy of Sciences; Tongji University; Beijing Technology and Business University（OPPO健康实验室; 广东欧珀移动通信有限公司; 北京大学; 西安电子科技大学; 中国科学院大学; 同济大学; 北京工商大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
PulseBound 是结合显式存储窗口信息边界的 PPG 表示学习器，通过存储后缀不变性等机制，在 MIMIC、VitalDB 等数据集上的预测性能优于基线模型，在下游任务中表现出色。
AI 中文摘要
从光体积描记法（PPG）进行的预测性表示学习，即便采用因果注意力机制，也可能违反因果信息访问规则，因为归一化、非局部变换或伴随视图可能依赖于未提供的样本。我们提出 PulseBound，一种结合生理结构化未来节拍预测与显式存储窗口信息边界的 PPG 表示学习器。与内容无关的截断将可见前缀与预测目标分开。仅前缀归一化、派生视图构建前的后缀替换以及对齐掩码确保编码器输入仅依赖于可见前缀和截断。这产生了存储后缀不变性：在模型状态、随机性、前缀和截断固定的情况下，改变存储后缀不会改变预测上下文。共享的水平条件头使用逐元素有效性掩码预测多达四个提取器验证的未来节拍的九个节律和形态描述符；可选的心电图（ECG）衍生的脉冲到达时间监督仅用于训练。在从 PulseBound 主干预训练中留出的 MIMIC 和 VitalDB 组上，PulseBound 相对于最后可见节拍持久性，将九状态变换空间平均绝对误差（MAE）分别降低了 28.06% 和 22.22%，在所有 40 个源-截断-水平单元的 MAE、MAE-Skill 和 Spearman 相关系数上均有提升。在 13 个下游任务的七个模型的单独比较中，PulseBound 在九个冻结线性探针和七个全微调任务上取得了最佳均值。存储后缀干预在审计精度下，在存储窗口接口下未导致预测上下文或预测的记录变化，且后缀输入梯度为零。这些发现区分了预测性生理表示学习的三个可测试方面：信息访问、监督未来结构和迁移。
英文摘要
Predictive representation learning from photoplethysmography (PPG) can violate causal information access even with causal attention, as normalization, nonlocal transforms, or companion views may depend on withheld samples. We introduce PulseBound, a PPG representation learner combining physiologically structured future-beat prediction with an explicit stored-window information boundary. A content-independent cutoff separates the visible prefix from the prediction target. Prefix-only normalization, suffix replacement before derived-view construction, and aligned masking ensure that encoder inputs depend only on the visible prefix and cutoff. This yields stored-suffix invariance: with fixed model state, randomness, prefix, and cutoff, changing the stored suffix cannot change the forecast context. A shared horizon-conditioned head predicts nine rhythm and morphology descriptors for up to four extractor-valid future beats, using elementwise validity masks; optional ECG-derived pulse-arrival-time supervision is restricted to training. On MIMIC and VitalDB groups held out from PulseBound backbone pretraining, PulseBound reduces nine-state transformed-space MAE relative to last-visible-beat persistence by 28.06% and 22.22%, respectively, with gains in MAE, MAE-Skill, and Spearman correlation across all 40 source-cutoff-horizon cells. In a separate comparison of seven models on 13 downstream tasks, PulseBound achieves the best mean on nine frozen linear-probe and seven full-fine-tuning tasks. Stored-suffix interventions cause zero recorded changes in forecast contexts or predictions, with zero suffix-input gradients at audited precision under the stored-window interface. These findings separate three testable aspects of predictive physiological representation learning: information access, supervised future structure, and transfer.
Commentsunder review
