This is an application-oriented course, addressed to beginners. Its target is to introduce AI technologies that can find successful application in structural engineering, including machine learning, neural networks, and deep learning, and to discuss, with real-world examples, how AI can be exploited to improve the design, construction, and maintenance of structures. The final goal is to provide students with the knowledge needed to identify, understand, and compare different deep learning techniques and analyze structural systems using appropriate techniques. The course also addresses potential challenges and ethical considerations in implementing AI in engineering. These may include issues related to data privacy and the need for regulations to ensure the safe and responsible use of AI. Finally, the course provides a step-by-step guide to implementing AI solutions in structural engineering, from identifying problems and setting objectives to evaluating and improving AI applications. Students will learn how to choose the right AI tools and technologies for their projects and how to implement AI solutions from analysis and design to maintenance.

By completing this course, the student will learn: - The intersection of AI and structural engineering: impact and benefits - AI technologies: Machine Learning, Deep Learning, Neural Networks - Applications of AI in structures: design optimization, analysis, structural health monitoring and predictive maintenance, digital twins - How to implement AI to structural systems: procedures, required data sets, training an AI model, specialized AI software for structural engineering.

Students must have a good knowledge of structural mechanics, linear algebra and attitude to problem analysis, as well as a basic knowledge of programming. Hands-on experience in using Python is an asset.

The course is organized into a theoretical and an applied module, developed in parallel throughout the semester. Theoretical module - Monte Carlo Methods - Linear and nonlinear regression models - Classification methods (k-Nearest Neighbors, Support Vector Machine, …) - Artificial Neural Networks - Convolutional Neural Networks - Data augmentation, generative adversarial learning - Physics-informed Machine Learning - Ethics Applied Module - Introduction to Python - Implementation of data-driven machine learning models - Sensor data processing with reference to real data sets - Information extraction - Structural optimization examples - Surrogate modeling for structural performance prediction - Monitoring and predictive maintenance - Constraining ML models with the laws of physics and engineering principles.

.

The course is divided into theoretical classes (27 h), exercise sessions with Python (15 h), analysis of case studies and applications (12 h), and 3-4 seminars held by software houses and professionals. Some exercises will be assigned as homework. For each of them, a short report is then required, to be delivered for evaluation during the exam.

- Dirk P. Kroese, Zdravko I. Botev, Thomas Taimre, Radislav Vaisman (2024) Data Science and Machine Learning. Mathematical and Statistical Methods. - Joel Grus (2015) Data Science from scratch. O’Reilly Media

Slides; Esercizi; Esercitazioni di laboratorio;

Modalita di esame: Prova orale obbligatoria; Elaborato scritto individuale;

Exam: Compulsory oral exam; Individual essay;

... The exam aims to assess the student ability to identify, understand, and compare different AI techniques to analyze structural problems. The exam consists of an oral exam, in which the student is asked to discuss the reports of the intermediate assignments, intended to ascertain the acquisition of practical skills in the use of the AI technologies. A report including all the assigned homework is to be delivered by the day of the oral exam. Furthermore, a few additional questions are asked to evaluate the comprehension of the background theory and the issues concerning the implementation/application of the AI techniques.

Gli studenti e le studentesse con disabilita o con Disturbi Specifici di Apprendimento (DSA), oltre alla segnalazione tramite procedura informatizzata, sono invitati a comunicare anche direttamente al/la docente titolare dell'insegnamento, con un preavviso non inferiore ad una settimana dall'avvio della sessione d'esame, gli strumenti compensativi concordati con l'Unita Special Needs, al fine di permettere al/la docente la declinazione piu idonea in riferimento alla specifica tipologia di esame.