EMG-FM-Bench: A Comprehensive Benchmark for Foundation Model Transfer and Adaptation on Electromyography

📄 arXiv: 2610.06450v1 📥 PDF

作者: Tianhao Wu, Xu Wu, Amirmohammad Radmehr, Jiawei Yu, Yi Wu, Phuc Nguyen, Jian Liu

分类: cs.LG

发布日期: 2026-10-05


💡 一句话要点

提出EMG-FM-Bench以解决肌电图模型迁移与适应性问题

🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)

关键词: 基础模型 肌电图 迁移学习 模型适应性 数据集整合 性能评估 生理信号处理

📋 核心要点

  1. 现有基础模型在肌电图(EMG)任务中的迁移性差异大,尤其在用户和任务变化时表现不佳。
  2. 提出EMG-FM-Bench基准,通过统一20个数据集,系统评估预训练模型在EMG任务中的迁移与适应性。
  3. 实验结果显示,五次适应能在70.2%的模型-数据集组合中提升宏F1分数,但未能完全恢复性能损失。

📝 摘要(中文)

基础模型(FMs)在时间序列和生理信号的开发中越来越普遍,但其在下游生理任务中的可迁移性仍不清楚。尤其在肌电图(EMG)中,由于信号分布在用户、传感配置、采集硬件和下游任务之间的显著差异,这一问题尤为复杂。本文介绍了EMG-FM-Bench,这是一个系统的基准,用于研究基础模型在EMG上的迁移和适应性。EMG-FM-Bench统一了20个公共数据集,包含超过100万个EMG片段,并评估了九个预训练基础模型在四个问题上的表现。结果表明,线性探测提供了关于预训练表示的有用信息,但完全微调可以显著改变下游EMG性能。预训练的好处在不同模型间差异显著,且在新用户上性能下降,但五次适应在70.2%的模型-数据集组合中改善了宏F1分数。

🔬 方法详解

问题定义:本文旨在解决基础模型在肌电图(EMG)任务中的迁移性和适应性问题。现有方法在不同用户和任务下表现不一致,缺乏系统性评估。

核心思路:通过构建EMG-FM-Bench基准,统一多个数据集,评估预训练模型在EMG任务中的表现,探索微调和用户变化对模型性能的影响。

技术框架:整体架构包括数据集整合、模型评估和性能分析三个主要模块。首先整合20个公共EMG数据集,然后对九个预训练模型进行评估,最后分析不同条件下的性能变化。

关键创新:EMG-FM-Bench的构建是本研究的核心创新,提供了一个系统的框架来评估基础模型在EMG任务中的迁移性,填补了现有研究的空白。

关键设计:在实验中采用了线性探测和完全微调两种策略,评估不同模型的预训练效果,关键参数设置和损失函数设计确保了实验的有效性和可比性。具体的模型架构和训练细节在论文中详细描述。

🖼️ 关键图片

fig_0
img_1
img_2

📊 实验亮点

实验结果显示,线性探测提供了有用的信息,而完全微调显著提高了下游EMG任务的性能。五次适应在70.2%的模型-数据集组合中改善了宏F1分数,但未能完全恢复性能损失,表明预训练的效果因模型而异。

🎯 应用场景

该研究的潜在应用领域包括医疗监测、康复训练和人机交互等。通过提高肌电图信号处理的准确性和适应性,能够为临床诊断和个性化治疗提供更有效的支持,具有重要的实际价值和未来影响。

📄 摘要(原文)

Foundation models (FMs) are increasingly being developed for general time series and physiological signals, yet their transferability to downstream physiological tasks remains poorly understood. This question is particularly challenging for electromyography (EMG), where signal distributions vary substantially across users, sensing configurations, acquisition hardware, and downstream tasks. We introduce EMG-FM-Bench, a systematic benchmark for studying foundation-model transfer and adaptation on EMG. EMG-FM-Bench unifies 20 public datasets with over 1 million EMG segments and evaluates nine pretrained foundation models across four questions: how pretrained models perform when frozen or fully fine-tuned, how much pretraining helps compared with training the same model from scratch, how well models generalize to new users with limited labeled data, and how performance changes across different EMG tasks. Across the benchmark, linear probing provides useful information about pretrained representations, but full fine-tuning can substantially change downstream EMG performance. Comparing each pretrained model with the same model trained from scratch shows that the benefit of pretraining varies substantially across models and is not universal. Performance decreases when models are evaluated on new users, while five-shot adaptation improves macro-F1 in 70.2% of evaluated model-dataset combinations but recovers only part of the lost performance. Model performance is highly consistent between upper- and lower-limb classification and remains strongly correlated with continuous EMG-to-text decoding. Together, these results provide a systematic view of when pretrained time-series models transfer effectively to EMG and how their performance depends on fine-tuning, user variation, and downstream task.