ActiLabel: A Combinatorial Transfer Learning Framework for Activity Recognition

TL;DR

ActiLabel introduces a novel transfer learning framework that leverages graph models to recognize human activities across diverse sensor devices, enhancing adaptability and accuracy in IoT environments.

Contribution

The paper presents a new combinatorial transfer learning framework, ActiLabel, which models structural similarities via dependency graphs to improve activity recognition across domains.

Findings

Outperforms state-of-the-art transfer learning methods

Effective across multiple public datasets

Demonstrates robustness to sensor diversity

Abstract

Sensor-based human activity recognition has become a critical component of many emerging applications ranging from behavioral medicine to gaming. However, an unprecedented increase in the diversity of sensor devices in the Internet-of-Things era has limited the adoption of activity recognition models for use across different domains. We propose ActiLabel a combinatorial framework that learns structural similarities among the events in an arbitrary domain and those of a different domain. The structural similarities are captured through a graph model, referred to as the it dependency graph, which abstracts details of activity patterns in low-level signal and feature space. The activity labels are then autonomously learned by finding an optimal tiered mapping between the dependency graphs. Extensive experiments based on three public datasets demonstrate the superiority of ActiLabel over state-of-the-art transfer learning and deep learning methods.