Acaba de ser aceite e publicado online, em formato aberto, um artigo da co-autoria de Manuel Parente, Head of AI no BUILT CoLAB, com o título “Predictive and prescriptive analytics in transportation geotechnics: Three case studies”.
Este artigo, que irá ser editado no volume 5 da revista “Transportation Engineering” da Elsevier em Setembro de 2021, tem os seguintes tópicos de pesquisa:
• Artificial intelligence (AI) can support Transformation Infrastructure Management (TIM);
• TIM predictive analytics can be obtained by using Machine Learning (ML);
• TIM prescriptive analytics can be obtained by using Evolutionary Computation (EC);
• Quality Neural Networks and Support Vector Machines predictions were obtained;
• Complex ML models can be opened by using eXplainable AI (XAI) methods.
Resumo do artigo:
«Transportation infrastructure is of paramount importance for any country. The construction, management and maintenance of this infrastructure is a complex task that requires a significant amount of resources (e.g., human work equipment, materials, maintenance costs). To better support this task, in the last decades several Artificial Intelligence (AI) data analysis tools have been proposed. In this paper, we summarize recent predictive and prescriptive AI applications to the transportation infrastructure field, underlying their strategic impact. In particular, we discuss three case studies: the design of better earthwork projects; the prediction of jet grouting soilcrete mechanical and physical properties (uniaxial compressive strength, stiffness and column diameter); and prediction of the stability level of engineered slopes.»
O artigo foi escrito em co-autoria com Joaquim Tinoco, António Gomes Correia, Paulo Cortez e David Toll, e está disponível para consulta no portal Science Direct aqui.