Teaching
Summer 2025: Learning from the Impossible (3 CP)
3 CP Seminar for Computer Science, Machine Learning, and Philosophy students. For more information see Alma.
This course is taught by Balthasar Grabmayr, Timo Freiesleben, and Sebastian Zezulka.
Introduction: Tuesday, 15.04., 2-4 pm c.t., Hörsaal TTR2, AI Research Building.
Fridays, 02.05 / 16.05. / 23.05. / 04.07., 09 am - 2 pm s.t., Hörsaal 1.3 Forum Scientarium.
At the core of mathematics, computer science, and machine learning lie formal results that show that certain things do not exist. For example,
Gödels incompleteness theorem shows that in any effective axiomatization of arithmetics, true statements exist that cannot be proven from the axioms.
The no-free lunch theorems show that no single learning algorithm is superior to all others across all possible environments.
The fairness impossibility theorems show that certain notions of formal fairness in predictions are generally incompatible, e.g. equalized odds and predictive parity.
These results are often thought to have profound philosophical and practical implications. But what exactly are these implications? How do they affect real-world applications, and might there be ways to navigate or mitigate the dilemmas they present?
In this seminar, we will try to get to the bottom of these questions. The seminar is divided into two blocks. In the first block, we (the instructors) will introduce you to the three impossibility results mentioned above. We will discuss their philosophical origins and provide accessible proofs to them. In the second block, you (the students) will present and discuss both the philosophical and practical implications of these theorems based on recent research articles.
This seminar is open to students from a wide range of disciplines, including philosophy, machine learning, computer science, and cognitive science. An openness to interdisciplinary collaboration, a willingness to engage with mathematical reasoning, and an interest in philosophical inquiry will enrich your experience in this course.
The seminar will be held in English.
Winter 2024/25: Philosophy of Science for Machine Learning (3 CP)
3 CP Seminar for Computer Science, Machine Learning, and Philosophy students. For more information see Alma.
15.10.2024 - 04.02.2025,
Tuesday, 2-4 pm c.t.
Lecture hall in the AI Research Building, Maria-von-Linden-Str. 6, Tübingen.
Summer 2024: Philosophy of Artificial Intelligence (3/6 CP)
Seminar jointly organized with Prof. Hong Yu Wong (Uni Tübingen).
See alma for more information.
Winter 2023/24: From Deep Learning to Rational Machines (3/6 CP)
Seminar organized together with Prof. Hong Yu Wong (Uni Tübingen) on Cameron J. Buckner's new book "From Deep Learning to Rational Machines" (2024). More information will follow.
Winter 2023/24: Philosophy of Science for Machine Learning (3 CP)
3 CP Seminar for Computer Science, Machine Learning, and Philosophy students. For more information see Alma.
Summer 2022: Ethics and Philosophy of Machine Learning (3 CP)
Taught together with Thomas Grote. You can find the Syllabus here.