Bio-PM 引导的合成 IMU 数据生成:在预训练编码器表示空间中做数据增强
Bio-PM-Guided Synthetic IMU Generation
做可穿戴传感或 HAR 的人可以看看这篇:不生成原始 IMU 信号,而是用 Bio-PM 编码器的 token 空间做扩散生成,仅靠合成数据训练就能拿到 0.68 的 TSTR macro-F1,加上增强后在 WISDM 上把 LOSO macro-F1 从 0.729 拉到 0.809。对比传统信号级增强,这种预训练表示空间生成路线更省数据,跨被试泛化是痛点的研究团队尤其该关注。
一项新研究探讨能否用预训练生物信号编码器 Bio-PM 生成的合成 IMU 数据,解决人体活动识别(HAR)中数据采集与标注昂贵、跨被试泛化差的问题。方法上的关键选择是不合成原始信号,而是用冻结的 Bio-PM 编码器把 10 秒 IMU 窗口转成 token 嵌入序列,再在该空间训练两类类条件解码器:一个自回归 transformer 解码器和一个基于去噪目标的扩散解码器,并另提出 PatchDecoder 做近似的原始波形重建。
评估采用分布性指标、嵌入空间可视化以及 leave-one-subject-out(LOSO)HAR 分类。结果显示扩散解码器在 Train-on-Synthetic-Test-on-Real(TSTR)设定下达到 0.68 的 macro-F1,说明生成的 token 足够真实,可用来训练能迁移到真实数据的分类器。把合成 token 增强与 320 维重力感知特征结合后,在 WISDM 基准 51 名被试上,LOSO macro-F1 从 0.729 提升到 0.809,且增益集中在动态相似、最难区分的活动类别上,具有统计显著性。
Bio-PM-Guided Synthetic IMU Generation
Collecting and labeling inertial measurement unit (IMU) data for human activity recognition (HAR) is expensive, and models trained on one group of subjects often fail to generalize to unseen individuals. Even with balanced data, classifiers struggle to distinguish dynamically similar activity classes and benefit from targeted augmentation. We explore whether synthetic IMU data generated in the representation space of a pretrained biosignal encoder, Bio-PM \cite{b1}, can serve as a practical augmentation source. Rather than synthesizing raw signals, we leverage Bio-PM's frozen encoder to convert 10-second IMU windows into token embedding sequences and train class-conditional decoders in this space: (1) an autoregressive transformer decoder that predicts tokens sequentially, and (2) a diffusion decoder that learns the token distribution through a denoising objective. We analyze generation realism through distributional metrics and embedding-space visualizations, evaluate downstream utility through leave-one-subject-out HAR classification, and present a PatchDecoder for approximate raw waveform reconstruction. Our diffusion decoder achieves a Train-on-Synthetic-Test-on-Real (TSTR) macro-F1 of 0.68, demonstrating that generated tokens are realistic enough to train a classifier that transfers to real data. Combining synthetic-token augmentation with a 320-dimensional gravity-aware feature representation raises LOSO macro-F1 from 0.729 to 0.809 over a real-only baseline across 51 subjects on the WISDM benchmark, with statistically significant gains concentrated in the most challenging activity classes. These results demonstrate that high-fidelity generation in a pretrained encoder's representation space is achievable and practically useful for cross-subject HAR augmentation.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include IMU, HAR, Bio-PM, diffusion models.
Source attribution: curated from arXiv cs.LG (@Deepali Singh, Deepesh Parpani, Aviral Gupta).