Research map / Time series / Foundation models for time series
Learning Fingerprints for Medical Time Series with Redundancy-Constrained Information Maximization
Huayu Li, ZhengXiao He, Xiwen Chen, Jingjing Wang, siyuan tian, Jinghao Wen, Ao Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add reconstruction pretraining (ablation) | TS-Fingerprint · ADFTD | accuracy: 48.83 -> 53.51 | Improved |
| add diversity loss to reconstruction pretraining (ablation) | TS-Fingerprint · ADFTD | F1: 50.48 -> 51.79 | Improved |
| add diversity loss to reconstruction pretraining (ablation) | TS-Fingerprint · ADFTD | AUROC: 70.21 -> 72.11 | Improved |
| add reconstruction pretraining (ablation) | TS-Fingerprint · PTB-XL | accuracy: 71.29 -> 73.56 | Improved |
| add diversity loss to reconstruction pretraining (ablation) | TS-Fingerprint · PTB-XL | F1: 62.66 -> 63.51 | Improved |
| add diversity loss to reconstruction pretraining (ablation) | TS-Fingerprint · PTB-XL | accuracy: 73.56 -> 73.98 | Improved |
| pretrain Ti-MAE then fine-tune (ablation) | Ti-MAE · ADFTD | F1: 45.88 -> 45.44 | No clear effect |
| pretrain SimMTM then fine-tune (ablation) | SimMTM · ADFTD | F1: 43.66 -> 45.59 | Improved |
| pretrain TS-Fingerprint then fine-tune (ablation) | TS-Fingerprint · ADFTD | F1: 45.80 -> 51.79 | Improved |
| use 8 instead of 6 fingerprint tokens (ablation) | TS-Fingerprint · PTB-XL | accuracy: 0.7071 -> 0.7398 | Improved |
| use 10 instead of 8 fingerprint tokens (ablation) | TS-Fingerprint · PTB-XL | accuracy: 0.7398 -> 0.6641 | Hurt |
| raise mask ratio from 0.6 to 0.8 at K=8 (ablation) | TS-Fingerprint · PTB-XL | accuracy: 0.7398 -> 0.7235 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Medformer — Medformer: A Multi-Granularity Patching Transformer for Medical Time-Series Classification (NeurIPS 2024) · better on 6 of 6 shared results
- iTransformer — iTransformer: Inverted Transformers Are Effective for Time Series Forecasting (ICLR 2024) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Medformer: A Multi-Granularity Patching Transformer for Medical Time-Series Classification (NeurIPS 2024 · similarity 0.88)
- Boosting Masked ECG-Text Auto-Encoders as Discriminative Learners (ICML 2025 · similarity 0.85)
- Guiding Masked Representation Learning to Capture Spatio-Temporal Relationship of Electrocardiogram (ICLR 2024 · similarity 0.85)
- Integrating Sequence and Image Modeling in Irregular Medical Time Series Through Self-Supervised Learning (AAAI 2025 · similarity 0.85)
- Parallel-Learning of Invariant and Tempo-variant Attributes of Single-Lead Cardiac Signals: PLITA (AAAI 2025 · similarity 0.85)
- From Token to Rhythm: A Multi-Scale Approach for ECG-Language Pretraining (ICML 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
