DISCOVER: Data-driven Identification of Sub-activities via Clustering and Visualization for Enhanced Activity Recognition in Smart Homes

2025-02-11

arXiv.org

AI Summary

In smart home environments, environment sensor-based human activity recognition (HAR) faces deployment bottlenecks including high annotation cost, reliance on pre-segmented sensor streams, and rigid activity granularity. To address these, we propose an unsupervised fine-grained sub-activity discovery framework. Our method integrates self-supervised feature learning, spectral clustering, and interpretable visualization (t-SNE/UMAP), enabling domain experts to achieve high-accuracy, semantically coherent activity relabeling by annotating only ~0.05% of the data—specifically, cluster centroids. Evaluated on benchmark HAR datasets, our approach automatically uncovers critical sub-activities overlooked by conventional coarse-grained labels, substantially improving label fidelity and model generalization. This significantly lowers practical deployment barriers while preserving scalability and interpretability.

Technology Category

• Machine Learning: Active Learning
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Computer Vision: Video Understanding & Activity Analysis

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Human Activity Recognition (HAR) using ambient sensors has great potential for practical applications, particularly in elder care and independent living. However, deploying HAR systems in real-world settings remains challenging due to the high cost of labeled data, the need for pre-segmented sensor streams, and the lack of flexibility in activity granularity. To address these limitations, we introduce DISCOVER, a method designed to discover fine-grained human sub-activities from unlabeled sensor data without relying on pre-segmentation. DISCOVER combines unsupervised feature extraction and clustering with a user-friendly visualization tool to streamline the labeling process. DISCOVER enables domain experts to efficiently annotate only a minimal set of representative cluster centroids, reducing the annotation workload to a small number of samples (0.05% of our dataset). We demonstrate DISCOVER's effectiveness through a re-annotation exercise on widely used HAR datasets, showing that it uncovers finer-grained activities and produces more nuanced annotations than traditional coarse labels. DISCOVER represents a step toward practical, deployable HAR systems that adapt to diverse real environments.

Problem

• Identifies fine-grained sub-activities from unlabeled sensor data
• Reduces annotation workload by clustering representative samples
• Enables flexible activity recognition without pre-segmented sensor streams

Innovation

• Unsupervised clustering discovers sub-activities from unlabeled data
• Visualization tool streamlines annotation of cluster centroids
• Minimal labeling reduces annotation workload to few samples