Robust Truth Discovery in Large-Scale Social Sensing for Cyber-Physical Systems

Abstract

This research addresses the challenge of extracting reliable information from massive, uncertain data streams in social sensing applications for cyber-physical systems (CPS). Existing truth discovery methods often struggle with scalability and the inherent uncertainty in social media data. We propose a novel scalable, uncertainty-aware framework that leverages advanced machine learning techniques to identify credible information sources and fuse their reports effectively. Our results demonstrate improved accuracy and efficiency compared to state-of-the-art methods, paving the way for more reliable CPS monitoring and control based on social sensing.

Introduction

Cyber-physical systems (CPS), such as smart grids and autonomous vehicles, rely increasingly on social sensing – leveraging information from diverse sources like social media and citizen sensors. However, this data is often noisy, incomplete, and conflicting, making truth discovery crucial. Current approaches frequently fail to handle the scale and uncertainty inherent in big data social sensing. Existing methods either lack scalability for massive datasets or struggle to effectively model and propagate uncertainty, leading to inaccurate insights and potentially compromising CPS operation. This research aims to address these limitations by developing a robust and scalable truth discovery framework specifically tailored for the unique challenges of CPS applications.

Objectives

• Develop a scalable algorithm for truth discovery from large-scale social sensing data in CPS applications.
• Implement an uncertainty-aware model to accurately represent and propagate uncertainty throughout the truth discovery process.
• Evaluate the proposed system's performance against existing state-of-the-art methods.