Enhancing Reliability and Trustworthiness in IoT Applications through Deep Learning-Based Data Imputation Techniques

Hakob Grigoryan


With the evolution of intelligent sensing devices and the Internet of Things (IoT), a vast amount of data is generated from various sources, including sensors, cameras, and network infrastructures, and is transmitted to servers for analysis. Data streaming from sensors in IoT systems might face quality issues like incompleteness due to sensor failures, network, or transmission errors. Addressing this problem is critical because when missing data is not handled properly, it results in inaccurate and unreliable analysis during the decision-making procedure. Traditional statistical methods for handling missing data, such as mean imputation or listwise deletion, are not suitable for IoT data due to their inability to adapt to dynamic streaming data. Despite the significant influence of data quality on decision-making systems, the impact of different data imputation techniques remains insufficiently investigated. The goal of my research visit to the University of Athens is to investigate the effectiveness of intelligent data-driven methodologies for the imputation task of IoT data streams. The research will be conducted over a three-month period, aiming to produce a comparative study that evaluates the capabilities of several Data Mining and Deep Learning algorithms for autonomously estimating missing values by exploiting temporal and spatial dependencies in sensor data. Through extensive experimentation and comprehensive assessment, I will present a collaborative high-quality research paper and an open-source model, summarizing research findings and promoting reproducibility. The project’s ultimate goal is to improve the reliability and trustworthiness of IoT applications by using advanced AI algorithms.

Keywords: Deep Learning, Artificial Intelligence, Data Mining, Internet of Things, missing data, data imputation, IoT, sensor data, Trustworthy AI, neural networks.

Scientific area: Artificial Intelligence, Internet of Things (IoT)

Bio: I am Senior Researcher specializing in AI, IoT, and Blockchain at NVISION Systems and Technologies S.L, based in Barcelona, Spain. I hold a Ph.D. in Machine Learning from Bucharest University of Economic Studies, a Master’s degree in Computer Science from Bucharest University, and a Bachelor’s degree in Cybernetics from National Polytechnic University of Armenia. With more than a decade of academic experience in data science and engineering, I have authored over 15 papers for both scientific journals and conferences. My research primarily concentrates on applying AI and Blockchain technologies across various sectors including IoT, finance, energy, and water. Currently, at NVISION, my technical development activities involve the design and implementation of decision support systems, predictive maintenance, and smart control mechanisms utilizing various AI and Data Mining algorithms.

Visiting period: May to August 2024 at National and Kapodistrian University of Athens