arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.13435cs.LG
收藏

PPG-Bench:面向PPG基础模型的统一基准

PPG-Bench: A Unified Benchmark for PPG Foundation Models

Mannas AI(曼纳斯人工智能)

Arvasu Kulkarni, Mahir Jain, Aditya Ray Mishra, Sandeep Singh, Siddharth Panwar

AI总结:

本研究提出PPG-Bench基准,评估5个PPG基础模型在9个数据集26项任务的表现,发现多数模型无显著优势,建议未来评估加入简单非学习型对照。

摘要:

光电容积描记(Photoplethysmography,PPG)正被可穿戴设备和临床设备以极大规模记录,越来越多的PPG基础模型(Foundation Models,FMs)被开发出来,用于为各种下游任务计算可复用的表征。这些模型难以比较:不同评估使用不同的数据集、任务定义、预处理方式、下游头模型、超参数调优方法和基线模型。我们提出PPG-Bench,这是一个统一的开源基准,用于评估5个公开的PPG基础模型,涉及9个开放获取数据集和26项任务,涵盖心血管状态估计、活动监测、睡眠分析、疾病筛查和身份识别。我们在4种适应性递增的下游机制下评估模型:线性探测、随机森林、冻结MLP探测和端到端微调,同时进行模型特定的超参数优化和受试者不相交交叉验证。我们还设置了3种对照:随机/恒定基线、确定性心率变异性(Pulse-Rate-Variability,PRV)特征集,以及通用时间序列基础模型(MOMENT)。在所有68种任务-机制组合中,35.3%的组合显示基础模型相比随机/恒定基线无显著优势,且在83.8%的组合中,没有PPG专用的基础模型同时优于PRV和MOMENT。我们还发现,增加下游头模型的容量通常会提升性能,但如果模型表现处于随机/恒定水平,可能会导致显著的性能下降。因此,当前的PPG基础模型尚不能声称能提供PPG信号的通用且可迁移的表征,该领域未来的评估应包含对简单非学习型对照的检验。基准的代码和评估更多模型的说明可在此处获取:https://

英文摘要:

Photoplethysmography (PPG) is increasingly recorded by wearable and clinical devices at enormous scale, and a growing number of PPG foundation models (FMs) have been developed to compute reusable representations for various downstream tasks. These models are difficult to compare: different evaluations use different datasets, task definitions, preprocessing, downstream heads, hyperparameter tuning, and baselines. We present PPG-Bench, a unifying open-source benchmark evaluating five public PPG FMs across nine open-access datasets and 26 tasks spanning cardiovascular state estimation, activity, sleep, screening, and identification. We evaluate models under four downstream regimes of increasing adaptability (linear probing, random forests, frozen MLP probes, and end-to-end finetuning), with model-specific hyperparameter optimization and subject-disjoint cross-validation. We also evaluate against three controls: random/constant baseline, a deterministic pulse-rate-variability (PRV) feature set, and a generic time-series FM (MOMENT). Across all 68 task-regime combinations, 35.3% show no meaningful FM advantage over the random/constant baseline, and in 83.8% no PPG-specific FM outperforms both PRV and MOMENT. We also find that added downstream head capacity generally improves performance, but can cause significant degradation if the model is performing at chance/constant level. Current PPG FMs therefore cannot yet claim to provide generalised and transferrable representations of PPG signals, and future evaluation in this field should include checks against simple, non-learned controls. Code for the benchmark and instructions to evaluate more models can be found here - https://github.com/mannas-ai/ppg-bench.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑