Abstract
In the context of the fourth industrial revolution, machine –learning based applications, such as 3D Printing, computer vision and robotics, are envisaged to digitize construction. This chapter introduces the reader to the basics of AI and focuses on the role of machine learning in the development of technological applications aimed at the improvement of the construction industry's efficiency and sustainability. It reviews the potential benefits of these applications and identifies factors like e.g. high costs, skills gap, dynamic and risky working environment that typically keep them at a very low practical usage in construction. Furthermore, a bibliometric search reveals that, despite its key role in the digitization vision, machine learning is rarely acknowledged in the relevant ‘Construction 4.0' literature. The chapter also identifies the cybersecurity issues emerging from machine learning applications' use and discusses mitigation strategies and technologies aimed at minimizing the impact of threats.
TopIntroduction
The Fourth Industrial Revolution, also referred to as “Industry 4.0”, is a recently introduced concept describing a disruptive innovation era, where the organizations and processes are digitally interconnected with the potential to reshape the value delivery mechanisms for services and products across the whole value chain (Bolpagni et al., 2022). There is no doubt that the current technological advancements have massively increased the potential quantity of information available, the resolution and frequency at which it is being captured, and the speed at which it can be processed, especially after the emergence of the Internet of Things (IoT), i.e., the radical evolution of internet into a network of interconnected objects with built-in sensors and computing capability (Konanahalli et al., 2022). As the volume and complexity of data available continuously grows, the learning abilities, the vast memory and the ever-increasing processing speed of computers, are more necessary than ever, to complement human intelligence. In this context, various subfields and applications of Artificial Intelligence (AI) such as machine learning, optimisation algorithms, natural language processing, robotics and computer vision can be applied to detect meaningful data patterns and support decision-making for complex real-world problems.
The disruptive innovation of Industry 4.0 has also brought an ‘intelligent construction’ era, widely reported as Construction 4.0, typified by technologies and applications which have been proven to effectively contribute to industrial operations’ efficiency such as the IoT, sensors, cloud computing and big data analytics. Furthermore, well-established in the industrial practice, machine –learning based applications, such as Additive Manufacturing (widely known as 3D Printing) and robots are starting to pervade the construction industry and attract the intense interest of construction management researchers. In this context, the limits between different scientific fields become increasingly diffuse and transform construction into an interdisciplinary industry (You & Feng, 2020; Forcael et al., 2020).
However, the use of AI in construction management, is not new at all. In the past few decades, dozens of relevant articles have used machine learning and optimization algorithms to tackle a very diverse range of issues, indicatively including the management of health and safety risks, the prevention of disputes between the stakeholders, the optimization of the site layout, the management of decision processes with regards to bidding, maintenance and resource allocation, among others. In the above context, this chapter will explore the role of machine learning in Construction 4.0 applications and will attempt to quantitatively and qualitatively assess the impact of the advent of Construction 4.0 in the AI-related research trends.
Besides the above, the literature shows consensus on the fact that beyond the extensive bibliometric analyses on potential benefits and opportunities (e.g. Oesterreich & Teuteberg 2016; Boton et al., 2021; Forcael et al., 2020; Abioye et al., 2021), the actual adoption of the Construction 4.0 technologies by the construction professionals is very low. The reasons for the construction industry struggling so much with the beneficial adoption of AI and other digital technologies has been attributed to a variety of factors including cultural, cost, skill, technological and data security barriers (Abioye et al., 2021). Regarding the latter, machine learning algorithms are indeed vulnerable to attacks that can transform them into a liability for the system they are intended to aid (Joseph et al., 2019). However, at the same time, machine learning has been highlighted as a promising means for intrusion detection, especially over IoT data (Shahin & Sabri, 2022). In this context, this chapter will also discuss the cybersecurity issues emerging from the use of machine learning applications as well as mitigation strategies, which aim to select proper learning methods to minimize the impact of threats and prevent failures. This is a topic of critical importance for the construction field as such a failure can potentially result not only in cost and time implications, but also lead to injuries or loss of life (Abioye et al., 2021).
Key Terms in this Chapter
Robots: Highly automated devices aimed at performing various tasks and capable of perceiving the environment through various sensors and machine learning -based control techniques.
Digital Twin: A digital representation of the physical and functional properties of a physical object or process that serves as its real-time digital counterpart.
Adversarial Learning: the field of machine learning which aims to study vulnerabilities of machine learning models and algorithms, and make them secure against adversarial manipulation.
Computer Vision: AI sub-field concerned with artificial simulation of the human sight, enabled by deep learning techniques and encompassing image classification, object detection and image semantic segmentation.
Cybersecurity: Cybersecurity is the practice of protecting critical systems and sensitive information from digital attacks.
Federated Learning: A machine learning approach where no data exchange is necessary, as it is based on local data samples, stored across multiple decentralized edge devices or servers.
Artificial Neural Networks: Machine learning models in which hypotheses take the form of complex algebraic circuits, typically organized into many layers, with tunable connection strengths.
Construction 4.0: Concept reflecting the vision for digital transformation of the construction industry, through the implementation of a wide range of emerging technologies and applications (e.g. cyber-physical systems, 3D Printing, robotics), in line with the forerunner paradigm of Industry 4.0 from the manufacturing sector.