Deep Auto-Set: A Deep Auto-Encoder-Set Network for Activity Recognition Using Wearables

TL;DR

This paper introduces Deep Auto-Set, a novel deep learning approach that models human activity recognition as a set prediction problem, leveraging auto-encoders for unsupervised feature learning to improve accuracy on sensor data.

Contribution

It presents the first deep neural network framework for activity set prediction in HAR, combining supervised set learning with unsupervised auto-encoder-based feature extraction.

Findings

Significant performance improvement over baseline models

Effective handling of multiple concurrent activities

Utilization of unlabeled data enhances recognition accuracy

Abstract

Automatic recognition of human activities from time-series sensor data (referred to as HAR) is a growing area of research in ubiquitous computing. Most recent research in the field adopts supervised deep learning paradigms to automate extraction of intrinsic features from raw signal inputs and addresses HAR as a multi-class classification problem where detecting a single activity class within the duration of a sensory data segment suffices. However, due to the innate diversity of human activities and their corresponding duration, no data segment is guaranteed to contain sensor recordings of a single activity type. In this paper, we express HAR more naturally as a set prediction problem where the predictions are sets of ongoing activity elements with unfixed and unknown cardinality. For the first time, we address this problem by presenting a novel HAR approach that learns to output activity sets using deep neural networks. Moreover, motivated by the limited availability of annotated HAR datasets as well as the unfortunate immaturity of existing unsupervised systems, we complement our supervised set learning scheme with a prior unsupervised feature learning process that adopts convolutional auto-encoders to exploit unlabeled data. The empirical experiments on two widely adopted HAR datasets demonstrate the substantial improvement of our proposed methodology over the baseline models.