PhilML2023: Philosophy of Science Meets Machine Learning

Machine learning methods have become a mainstay in the tool-kit of various scientific disciplines. PhilML’23 offers an opportunity to explore whether and how exactly recent developments in the field of machine learning potentially transform the process of scientific inquiry. For this purpose, it sets out to analyse the field of machine learning through the lens of philosophy of science, including cognate fields such as epistemology and ethics. In addition, we are also interested in contributions from machine learning researchers/scientists, addressing foundational issues of their research. Similar to the previous workshops, we try to bring together philosophers from different backgrounds (from formal epistemology to the study of the social dimensions of science) and machine learning researchers.

The workshop`s central topics are:

(i) A critical reflection on key-concepts, such as ‘learning’, ‘causal inference’, ‘robustness’, ‘explanation’ or ‘understanding’.

(ii) The implications of machine learning for the special sciences, e.g. cognitive science, biology, social science or medicine.

(iii) The ethics of machine learning-driven science, e.g. the moral responsibilities of researchers, ethical issues in model evaluation, or issues at the intersection of science and policy.

(iv) Social aspects of machine learning-driven science, e.g. the impact of funding structures on research.

Invited Speakers

Florian Boge (Dortmund)
Frederik Eberhardt (Caltech)
Marta Halina (Cambridge)
Atoosa Kasirzadeh (Edinburgh)
Michael Knaus (Tübingen)
Silvia Milano (Exeter)
Emma Pierson (Cornell) -- online
Tom Sterkenburg (LMU)
Mariya Toneva (MPI-INF)
Claire Vernade (Tübingen)
Kino Zhao (Simon Fraser)

Each abstract will be reviewed by multiple philosophers/ML researchers from the Tübingen community. Selection criteria are research quality, novelty, and topical diversity. The workshop is organised by the Cluster of Excellence ‘Machine Learning: New Perspectives for Science’ at the University of Tübingen. For questions, please email: [email protected]

The call for abstracts is opened. Please submit an anonymised extended abstract (750 words not including references), along with a cover sheet containing your name, email-address and institutional affiliation until April 21 to

[email protected]

The final decisions will be announced by mid-May