Embedded Inter-Subject Variability in Adversarial Learning for Inertial Sensor-Based Human Activity Recognition

TL;DR

Results indicate that the proposed adversarial task effectively reduces inter-subject variability among different users in the feature space, and it outperforms adversarial tasks from previous works when integrated into the framework.

Abstract

This paper addresses the problem of Human Activity Recognition (HAR) using data from wearable inertial sensors. An important challenge in HAR is the model's generalization capabilities to new unseen individuals due to inter-subject variability, i.e., the same activity is performed differently by different individuals. To address this problem, we propose a novel deep adversarial framework that integrates the concept of inter-subject variability in the adversarial task, thereby encouraging subject-invariant feature representations and enhancing the classification performance in the HAR problem. Our approach outperforms previous methods in three well-established HAR datasets using a leave-one-subject-out (LOSO) cross-validation. Further results indicate that our proposed adversarial task effectively reduces inter-subject variability among different users in the feature space, and it outperforms adversarial tasks from previous works when integrated into our framework. Code: https://github.com/FranciscoCalatrava/EmbeddedSubjectVariability.git

Figures (3)

• Figure 1: Overview of our framework: Steps 1–2 pre-train all module weights. Step 3 introduces adversarial learning to minimize inter-subject variability. Evaluation is performed on previously unseen users. Data distribution $A$ is used for the activity classification (Eq. \ref{['eq:Adataset']}), while $A'$ is the data distribution used in the adversarial task (Eq. \ref{['eq:A_prime_definition']}). The circular arrow in Step 3 illustrates the iterative nature of adversarial learning.
• Figure 2: First column: $F1-Score_{M}$ for MHEALTH, PAMAP2, and REALDISP (top-down). Second column: Percentage change in Wasserstein distance from Step 2 to Step 3 between training and test set distributions. The top plot gives the distance change per dataset, while the middle (PAMAP2) and bottom (REALDISP) plots detail the distance change per activity. Positive bars (blue) mark a reduction (distance after Step 3 is smaller than after Step 2); negative bars (red) mark an increase.
• Figure 3: Average accuracy and $F1-Score_M$ for variations in $w_A$ (top), $w_R$ (middle), and $w_C$ (bottom) for PAMAP2.