A Light-weight Deep Human Activity Recognition Algorithm Using Multi-knowledge Distillation

TL;DR

This paper introduces SMLDist, a lightweight deep learning model for human activity recognition that leverages multi-knowledge distillation and auto-search to achieve high accuracy while maintaining energy and computation efficiency on resource-limited devices.

Contribution

The paper proposes a novel multi-level knowledge distillation framework with an auto-search mechanism, improving HAR accuracy and robustness with reduced model size and computational cost.

Findings

SMLDist outperforms state-of-the-art HAR models in accuracy and F1 score.

The model is energy-efficient and suitable for deployment on resource-limited platforms.

It demonstrates strong generalization across multiple datasets.

Abstract

Inertial sensor-based human activity recognition (HAR) is the base of many human-centered mobile applications. Deep learning-based fine-grained HAR models enable accurate classification in various complex application scenarios. Nevertheless, the large storage and computational overhead of the existing fine-grained deep HAR models hinder their widespread deployment on resource-limited platforms. Inspired by the knowledge distillation's reasonable model compression and potential performance improvement capability, we design a multi-level HAR modeling pipeline called Stage-Logits-Memory Distillation (SMLDist) based on the widely-used MobileNet. By paying more attention to the frequency-related features during the distillation process, the SMLDist improves the HAR classification robustness of the students. We also propose an auto-search mechanism in the heterogeneous classifiers to improve classification performance. Extensive simulation results demonstrate that SMLDist outperforms various state-of-the-art HAR frameworks in accuracy and F1 macro score. The practical evaluation of the Jetson Xavier AGX platform shows that the SMLDist model is both energy-efficient and computation-efficient. These experiments validate the reasonable balance between the robustness and efficiency of the proposed model. The comparative experiments of knowledge distillation on six public datasets also demonstrate that the SMLDist outperforms other advanced knowledge distillation methods of students' performance, which verifies the good generalization of the SMLDist on other classification tasks, including but not limited to HAR.