# Learning Fingerprints for Medical Time Series with Redundancy-Constrained Information Maximization Source: [Learning Fingerprints for Medical Time Series with Redundancy-Constrained Information Maximization]() Collection index: [llms.txt]() Data updated: 2026-05-05T05:00:04.460Z Authors: Huayu Li, ZhengXiao He, Xiwen Chen, Jingjing Wang, Siyuan Tian, Jinghao Wen, Ao Li Published: 2026-04-30T18:33:40.000Z ## Abstract Learning meaningful representations from medical time series (MedTS) such as ECG or EEG signals is a critical challenge. These signals are often high-dimensional, variable-length and rife with noise. Existing self-supervised approaches, such as Masked Autoencoders (MAEs) are highly effective for pre-training general-purpose encoders. However, they do not explicitly learn compact and semantically interpretable latent representations, typically relying on heuristic aggregation strategies such as global average pooling or a designated \[CLS\] token. We propose a novel framework that compresses a variable-length MedTS into a fixed-size set of $k$ latent Fingerprint Tokens. Our architecture employs a cross-attention bottleneck to generate these tokens and is trained with a dual-objective function. The first objective is a reconstruction loss, which ensures the tokens are \\textit{sufficient statistics} for the original data. The second, a diversity penalty based on the Total Coding Rate (TCR), explicitly minimizes the redundancy between tokens, encouraging them to become statistically \\textit{disentangled} representations. We present the theoretical justification for our method, framing it as a novel \\textbf{Disentangled Rate-Distortion} problem. This approach produces a low-dimensional, interpretable, and sample-efficient representation, where each token is encouraged to capture an independent factor of variation, paving the way for more robust digital biomarkers. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Medical Time Series Classification on PTB 2-Classes](): Medical Time Series Classification; PTB 2-Classes; Accuracy: 85.39. - [Medical Time Series Classification on APAVA 2-Classes](): Medical Time Series Classification; APAVA 2-Classes; Accuracy: 80.29. - [Medical Time Series Classification on SleepEDF 5-Classes](): Medical Time Series Classification; SleepEDF 5-Classes; Accuracy: 84.18. - [Medical Time Series Classification on FLAAP 10-Classes](): Medical Time Series Classification; FLAAP 10-Classes; Accuracy: 77.59. - [Medical Time Series Classification on UCI-HAR 6-Classes](): Medical Time Series Classification; UCI-HAR 6-Classes; Accuracy: 90.87. - [Medical Time Series Classification on PTB-XL 5-Classes](): Medical Time Series Classification; PTB-XL 5-Classes; Accuracy: 73.98. - [Medical Time Series Classification on ADFTD 3-Classes](): Medical Time Series Classification; ADFTD 3-Classes; Accuracy (%): 53.91. ## Original sources and implementation [Original paper]() [PDF]()