Key points are not available for this paper at this time.
Infrastructure systems are the backbone of modern society, and their timely recovery following natural hazards is vital for safeguarding human well-being and economic stability. However, the recovery process is inherently complex and involves dynamic interactions between multiple sectors, uncertainties in decision-making, and constraints in resource allocation. As artificial intelligence (AI) continues to advance, machine learning (ML) has emerged as a powerful approach for enhancing understanding and supporting data-driven decisions in infrastructure recovery. Despite the growing interest, existing studies on ML applications in this context remain fragmented. To address this gap, this study conducted a systematic review of the literature on ML applications in infrastructure recovery from natural hazards, following the guidelines of Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA). The review identifies key application areas, categorizes ML techniques across learning paradigms, and highlights common challenges. By synthesizing current developments, this work aims to provide a comprehensive overview of the field and offer insights for researchers and practitioners seeking to leverage ML for more effective, adaptive, and resilient recovery strategies. It also outlines future directions to advance ML integration in this emerging domain and promote its broader application in disaster recovery research and practice.
Li et al. (Thu,) studied this question.