跳到正文
arXiv cs.CL· Elif Dorukba\c{s}{\i}, \.Ilker T\"urker·· 23 小时前AI 评分31

面向大语言模型的紧凑可解码心电图文本编码

Compact and Decodable Electrocardiogram Text for Large Language Models

AI 导读

一种确定性编码将十二导联心电图中的八个独立导联写成振幅字母与精确时间间隔,其余四个肢体导联通过代数关系恢复,在 PTB-XL 数据库 21,388 条记录上全部压缩至 4,080 tokens 以内(平均 2,866 tokens,较逐样本转录减少 64.3%),且导联字符串可精确还原。

正文 · 原文

View PDF HTML (experimental)

Abstract:When a twelve-lead electrocardiogram is written as text for a large language model, the number of tokens has to be balanced against the information lost through compression. This paper presents a deterministic encoding that writes eight independent leads as amplitude letters and exact time gaps and recovers the remaining four limb leads algebraically. The key points are selected with a cardiac-cycle-dependent tolerance, the amplitudes are parameterized by a 26-letter logarithmic codebook learned from a training set, and a 10-step tolerance schedule keeps all records within a fixed token budget. The encoding was applied to 21,388 records from the public PTB-XL database and evaluated using an exact decode-and-re-encode round trip, waveform reconstruction error, and a paired comparison of convolutional classifiers trained with 5 different seeds on signals from the original and decoded databases. All records were contained in 4,080 tokens (average: 2,866 tokens, 64.3% fewer than a per-sample transcription), and all stored lead strings were exactly reproduced. The average correlation between the decoded and reference signals was 0.982. The macro-averaged area under the receiver operating characteristic curve was 0.9167 for the reference signals and 0.9157 for the decoded signals (paired difference -0.0010; 95% patient-level bootstrap interval, -0.0028 to 0.0007), within the pre-specified non-inferiority margin of -0.010; this is not a confirmatory result, because the test fold had been used during encoder development. The diagnostic content of a compact text can therefore be preserved, while the language model's ability to read the text is yet another matter.
Comments: 9 pages, 5 figures, 5 tables. Preprint of a chapter submitted to "Advances in Artificial Intelligence and Data Science: Methods, Systems, and Applications" (Karabük University Press). Code: this https URL
Subjects: Computation and Language (cs.CL)
ACM classes: I.5.4; J.3
Cite as: arXiv:2611.22827 [cs.CL]
  (or arXiv:2611.22827v1 [cs.CL] for this version)
  https://doi.org/10.48550/arXiv.2611.22827

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Elif Dorukbaşı [view email]
[v1] Tue, 17 Nov 2026 14:13:27 UTC (902 KB)

来源:arXiv cs.CL · arxiv.org