Prof. Dr. Michael A. Kraus, M.Sc.(hons)

Scientific Machine and Deep Learning for Design and Construction in Civil Engineering

101-0139-00L, Time: Mondays 14:00 – 18:00 in HPK D3 online via Zoom; Office hours: Monday: 1 - 2 pm

Course Description

In recent years, the Artificial Intelligence technology was introduced to the Architecture, Engineering and Construction (AEC) industry. Numerous application achievments have since been reached in this interdisciplinary field. However, yet today fundamental questions on the application of the relevant AI subgroups of Machine and Deep Learning for design, verification and construction situations have not answered. This lecture wants to equip current students as practicioners and researcher of the future with a basic yet solid understanding of the AI technology and what essential points to consider when applying AI to this specific domain.

The availability of data together from different sources together with the interoperability of modern Building Information Modeling (BIM) software within the design and verification workflow for AEC allow applicability of the new AI technology. This augmented building project delivery mode leverages an increase in efficiency and decreases risks due to lack of data-informed decisions in this trillion-dollar industry. However, cautious and well-thought steps need to be taken in the right direction, in order for such technologies to thrive in an industry that showcases inertia in technological adoption. To that end, this lecutre sheds some ligh on trustworthy and explainable AI. The hands-on approach during exercieses but also by conducting projects with real world applications of AI in different fields of AEC guarantees tangible outcomes of this lecture for the participants.

Course Structure

The course will be delivered as a three phase lecture:

  1. The first phase introduces fundamentals of the Machine and Deep Learning (ML/DL) technology, as building blocks to be considered during the development of related applications. Highlight of this lecture part are the various methods for deploying civil engineering domain knowledge to ML/DL to form Scientific ML (SciML). The approaches are discussed concerning latest developments and implementation obstacles.
  2. The second phase consists of 2 guest lectures with specific state-of-the-art industry and research applications of SciML.
  3. The third phase consists of classes for guiding and supporting participants in solving and deploying their project.

In the points of connection, students will see the importance of taking into account the application requirements when designing an AI system, as well as their impact on the building blocks. Guest speakers from both, the AI and AEC domains, provide complementary impulses for participants of this lectures.

Course Objectives

This course will present methods of scientific machine and deep learning (ML / DL) for applications in design and construction in civil engineering. After providing proper background on ML and the scientific ML (SciML) track, several applications of SciML together with their computational implementation during the design and construction process of the built environment are examined.

By the end of the course students will develop computational thinking related to the combination of domain knowledge and latest computer science AI technology for scientific machine learning applications within the AEC domain. Specifically, the students will:

Instructors

Michael Kraus
Dr. Michael A. Kraus, M.Sc.(hons)
PostDoctoral / Senior Researcher in SciML4AEC and Co-Leader of the Immersive Design Lab of Design++)
Instructor and Lecturer

Danielle Griego
Dr. Danielle Griego
Executive Director of Design++ and PostDoctoral Researcher
Co-{Instrucor ; Lecturer}

Course Schedule

(Subject to change)

DATE CLASS TOPIC MATERIAL
27.09 Introductory Class slides
27.09 Fundamentals of Machine Learning - Part 1: Data and essential Maths/Statistics slides slides video
04.10 Fundamentals of Machine Learning - Part 2: Supervised Learning: Basics slides video
04.10 Exercise 1: Introduction to Python and Pandas slides/notebook
11.10 Fundamentals of Machine Learning - Part 3: Supervised Learning: Classification and Regression slides slides video
18.10 Fundamentals of Machine Learning - Part 4: Data Processing and Visualisation slides video
18.10 Student Projects Pitches  
18.10 Exercise 2: Data Processing and Visualisation slides/notebook
25.10 Fundamentals of Machine Learning - Part 5: Unsupervised Learning and Optimization Details slides slides video
25.10 Exercise 3: ML Workflow, Regression and Classification slides/notebook
01.11 Introduction to Deep Learning slides slides video
01.11 Exercise 4: Unsupervised Machine Learning slides/notebook
08.11 1st Project Consultation (in person, at ETH Hönngerberg)  
15.11 Scientific Machine and Deep Learning video
22.11 Guest Talk, Graph-NeuralNetwork based SciML, by Professor Julija Zavadlav, Dept. of Mechanical Engineering, TU Munich video
22.11 Scientific Machine and Deep Learning slides/notebook video
22.11 Exercise 5: Feature Engineering slides/notebook
29.11 2nd Project Consultation (in person, at ETH Hönngerberg)  
06.12 Guest Talk, Physics-informed Neural Networks at scale, by Mohammad Nabian, NVIDIA  
13.12 Final Project Presentation (in person and online, at ETH Hönngerberg)  
20.12 Hand-in of Final Project Report (online / email))  

Course Evaluation