arXiv cs.AI· Tsai-Min Chen (Research Center for Information Technology Innovation, Academia Sinica, Taipei, Taiwan, RIKEN Center for Interdisciplinary Theoretical and Mathematical Sciences), Jun Seita (RIKEN Center for Interdisciplinary Theoretical and Mathematical Sciences)·· 56 分钟前AI 评分31
TIBECG:面向心电波形发布中导联特定人口隐私的时间信息瓶颈方法
TIBECG: Temporal Information Bottleneck for Lead-Specific Demographic Privacy in ECG Waveform Release
AI 导读
中研院与 RIKEN 研究团队提出 TIBECG,一种用于心电(ECG)波形发布的人口属性隐私保护架构时间信息瓶颈方法。该 2,302,337 参数的处理器无需年龄或性别标签即可拟合,在 6,386 条记录、十折、12 导联和 21 种隐变量长度下,各导联干净 macro-F1 至少保留 95%,年龄/性别区间增益较最佳常规变换中位数高出 22.08 个百分点。
正文 · 原文
Authors:Tsai-Min Chen (1 and 2), Jun Seita (2) ((1) Research Center for Information Technology Innovation, Academia Sinica, Taipei, Taiwan, (2) RIKEN Center for Interdisciplinary Theoretical and Mathematical Sciences (iTHEMS), RIKEN, Tokyo, Japan)
Abstract:Transmitting an electrocardiogram (ECG) waveform can expose demographic information beyond an intended diagnostic task. We evaluate TIBECG, an architectural temporal information bottleneck, on individual leads extracted from hospital twelve-lead recordings. A 2,302,337-parameter processor varies its latent length while retaining the same nine-score consistency objective, without age or sex labels in protector fitting. Across 6,386 recordings, ten folds, 12 leads and 21 lengths, retrospective selections retained at least 95% of each lead's clean macro-F1. Their weaker age/sex interval gain exceeded the best eligible conventional transformation on all leads, by a median 22.08 percentage points. The implemented sampling-rate comparison shows that direct downsampling usually suppresses demographic inference more strongly but at a larger loss of diagnostic utility, whereas TIBECG favors utility preservation. Selection stability also varies by lead. Lower latent Kullback-Leibler (KL) divergence does not reliably rank residual sex leakage at a fixed length. These findings support lead-specific evaluation of waveform processing, not validated consumer deployment or expert-readable reconstruction. When the waveform-input diagnostic classifier is hosted as part of a proprietary cloud service rather than deployed on the device, waveform transmission is required to obtain its scores.
| Comments: | 12 pages, 9 figures, 2 tables. Corresponding author: Tsai-Min Chen |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Signal Processing (eess.SP) |
| Report number: | RIKEN-iTHEMS-Report-26 |
| Cite as: | arXiv:2611.21694 [cs.LG] |
| (or arXiv:2611.21694v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.21694 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Tsai-Min Chen Ph.D. [view email]
[v1]
Mon, 16 Nov 2026 11:24:17 UTC (2,419 KB)
来源:arXiv cs.AI · arxiv.org