The rise of data capture and storage capabilities has enabled greater data granularity and the sharing of data sets. This broader shift to big data requires effective ways to process and extract value from information, aided by methodological advances in computer science and improvements in hardware capabilities. Machine learning (ML), a subset of artificial intelligence, typically processes input parameters using algorithms to generate outputs. Within geotechnical engineering, several areas have made progress in utilizing ML to automate or optimize various procedures. These can be broadly categorized into three overarching topics: material and site characterization, evaluation of system performance, and geotechnical earthquake engineering. Each topic requires its own ML implementation approach, with progress evident across all areas. As the intersection between ML and geotechnical engineering is not yet fully formalized, this is an opportune moment to evaluate current developments. This paper presents a literature review of ML approaches in geotechnical engineering. The paper highlights and critically summarizes progress within the three topics, addresses current challenges, and identifies opportunities to pursue in the coming years.
Cheng et al. (Thu,) studied this question.