用卷积替代循环结构，让可穿戴设备上的人体动作识别更省电更快。
LITEWAY: LIghtweight HAR via Temporal Efficient highWAY
- 用结构化卷积分解代替循环网络，实现全卷积的时序建模。
- 模型大小缩小4.06到9.52倍，能耗降低1.46到3.14倍。
- 适合资源受限设备，尤其适合低功耗可穿戴应用。
可穿戴设备上的人体活动识别（HAR）受深度学习模型计算与能耗限制，现有轻量级方法多依赖循环结构（如GRU、LSTM），影响并行性且增加推理延迟。本文提出LITEWAY，一种模态无关的全卷积框架，通过结构化卷积分解替代循环时序建模，结合轻量卷积块、步进式时间处理和卷积注意力池化，高效捕捉时序依赖的同时降低计算复杂度。在16个HAR数据集上对比TinyHAR、TinierHAR和MLP-HAR，LITEWAY在保持竞争力宏F1值的同时，相比这些模型，轻量版（Light）模型尺寸缩小4.06–9.52倍，完整版（Full）缩小3.87–9.07倍；部署实验显示，相比TinierHAR和MLP-HAR，能耗降低2.29–3.14倍（Light）和1.46–2.01倍（Full）。代码已开源。
原文摘要 · Abstract (English)
Wearable human activity recognition (HAR) remains challenging due to the computational and energy constraints of deep learning models on resource-limited devices. Existing lightweight approaches often rely on recurrent architectures (e.g., GRU and LSTM), limiting parallelism and increasing inference latency. We propose LITEWAY, a modality-agnostic, fully convolutional framework for multichannel sensor time series that replaces recurrent temporal modeling with structured convolutional decomposition. LITEWAY combines lightweight convolutional blocks, strided temporal processing, and convolution-attention pooling to efficiently capture temporal dependencies while reducing computational complexity. We evaluate LITEWAY on 16 HAR datasets against TinyHAR, TinierHAR, and MLP-HAR. LITEWAY achieves competitive macro F1 while reducing model size by 4.06x-9.52x (Light) and 3.87x-9.07x (Full) compared with TinyHAR and TinierHAR. Deployment experiments further show energy reductions of 2.29x-3.14x (Light) and 1.46x-2.01x (Full) compared with TinierHAR and MLP-HAR, highlighting efficient fully convolutional temporal modeling for wearable HAR. The source code is publicly available at https://github.com/dominique-nshimyimana/liteway.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。