Artificial intelligence

For EPA and COSEM students

System Theory, Object Orientation, Discrete Event System Specification, Multi-Formalism Simulation, Distributed Simulation, Real-Time Simulation, Multi-Paradigm Simulation, Multi-Resolution Simulation, and Simulation Language selection will be the core topics of the course. After an introduction to system theory, the inner working of simulation environments will be illustrated on the basis of the DEVS, DESS, and DTSS formalisms. Then, possible integrating of the different formalisms will be shown. Several special topics will be taught, based on the latest research in simulation. This material will be illustrated in intensive and interactive courses. In addition to the lecture topics, several other simulation topics will be studied by groups of students, who will write a scientific paper, and present their findings in class. These topics can be focused on the MSc program that the students participate in; special topics to study are available for TIL, CoSEM, EPA, Computer Science, and other students.  Finally, groups of students will study a simulation package in-depth and discuss the commonalities and differences with other packages. Again, the package chosen can be targeted at the MSc program of the students. TIL students can, e.g. study a package that is more aimed at logistics and transport, CoSEM students can focus on a package that is used in systems design, whereas EPA students can focus on a package used in policy analysis. Most packages can be made available for use from home. The course contains very little overlap with other simulation courses such as EPA1321, EPA1351, TB233A. It requires experience with a simulation environment such as Arena, Simio, or Toma.

For EPA students and COSEM students

Machine Learning (ML) is rapidly changing the way we do science and engineer systems and services, both inside and outside academia. Due to its increasing and widespread use, ML will be more and more engrained in, and be an integral part of, future societies. Successful adoption of ML by societies in all its breadth does not only require skilled ML professionals that do the hard-core programming (i.e. professionals with full-fledged computer science backgrounds), but also of the sort that have profound domain knowledge where ML is applied, such as in sociotechnical systems. ML will increasingly be a part of the puzzle to solve complex challenges of today’s networked, urbanized knowledge societies. Therefore, TPM MSc students of all programmes must have a thorough understanding of ML. 

For MOT students

For MOT Students

The course first provides an introduction to theories of normative ethics: utilitarianism, deontology, virtue ethics. The distinction between descriptive and normative claim, epistemology and ethics, values, norms, and virtues.

• Ethical and epistemological issues in AI: identification of six different sources of ethical and epistemological concerns in AI.
• Bias in AI: Automated-decision making systems are polluted with different forms of bias that can be found at a different level. Here we address the different forms of bias and what can be done about it.
• Explanatory AI and the right to explanation: The GDPR grants the so-called "right to an explanation." But what does this means, and how far have we got into explanatory AI? In this unit, we explore the scope and limits of current approaches to explainable AI and how this affects users' rights
• Privacy: there are different forms in which privacy is studied and understood. In addition, this unit discusses the value of privacy in AI in comparison with other values, such as explainability, accuracy, bias, and fairness)
• Trustworthy AI: What does it mean to trust an AI system? is there more than one way to trustworthy AI? can we design trust? In this unit, we will address questions such as these. We will also present and discuss the two major approaches to trustworthy AI, namely, transparency and computational reliabilism. Lights and shadows of both. 

The course will cover various concepts of AI governance. It will lean towards combining concepts that center marginalized communities and address issues of oppression and liberation, drawing from cybernetics, (eco)systems theory, critical algorithm studies, intersectionality studies, reflection and speculative and critical design for community empowerment. It will likely look at different kinds of ecosystems in which algorithms/data/AI play a role, including online ecosystems, work and labor, energy systems and ecological systems. The idea is to have teams of students study a particular kind of ecosystem and bring to bear a mix of the above contexts to better assess the implications and potential roles of data and algorithms and to identify ways of designing for social movement and change.