The third workshop on Metascience for Machine Learning
5 Dec 2025, Inspiration from other fields -- Philosophy and Psychology.
The workshop is kindly sponsored by ELLIS unit Delft.
Our third workshop is on Friday, December 5th, 2025 from 10:00-15:00; register here; tentative schedule, and location at Firma van Buiten (address: Thijsseweg 1, 2629 JA Delft).
Schedule
10:00: Walk in with coffee/tea
10:10: Welcome presentation
10:20: Presentation by Abigail Nives Delgado (University of Utrecht), talk + discussion
11:15: Presentation by Daniel Lakens (TU Eindhoven), talk + discussion
12:10: Lunch
13:00: Brief update subgroup activity (based on prior reporting)
13:10: Plenary discussion on new topics/activities/directions
13:30: Break out discussions
14:15: Plenary summary
14:35: Closing
14:45: End
Keynote speaker presentations:
Invited talk by Abigail Nives Delgado (University of Utrecht)
Title: Philosophy of Science as a Tool for Scientific Practice
Abstract: This talk introduces philosophy of science and its relevance for scientific and technological practice. It focuses on how examining the epistemic and non-epistemic values that shape research—such as accuracy, simplicity, social relevance, or ethical concerns—can help reveal implicit assumptions and biases in methods, models, and theories. Drawing on philosophy of science in practice, feminist philosophy of science, and integrated history and philosophy of science, the talk shows how philosophical analysis can improve critical reflection on scientific work. Through concrete examples, it illustrates how these approaches can support better research design, interpretation, and decision-making in science and technology.
Bio: Abigail Nieves Delgado is associate professor at the Freudenthal Institute, Utrecht University. She works on the history and philosophy of the life sciences, especially on racialization practices in the history of science as well as in contemporary biomedical research (e.g., in microbiome research and epigenetics) and in biometric technologies (e.g., facial recognition). She also investigates the politics of transdisciplinary knowledge production and the history of ethnobiology in Latin America. She is the principal investigator of the project “Microbiome research and race in the ‘Local South’” funded by the NWO. This project aims to develop epistemically fruitful, non-discriminatory, and locally relevant classification criteria of human diversity to be used in human microbiome research instead of pernicious and ambiguous categories of race.
Invited talk by Daniel Lakens (TU Eindhoven)
Title: How to Improve Our Science: Lessons from the Replication Crisis in Psychology
Abstract: Psychological Science has gone through two crises of confidence. The first started in the 1960s and lasted until the end of the 1970s, and the second crisis dominated the 2010s. In both these crises researchers discussed fundamental concerns about the replicability of findings, the strength of theories in the field, the societal relevance of research, the generalizability of effects, and problematic methodological and statistical practices. On the basis of extensive quotes drawn from articles published during both crises, I explore the challenges our field faced, which are not specific to psychology, but relevant for many disciplines in science. I then review recent Open Science developments in our field aimed at addressing these challenges, and reflect on whether and why they have been successful, or not, as well as on the relevance of these developments for other scientific fields.
Bio: Daniel Lakens is an Associate Professor of Metascience and chair of the Ethical Review Board at the Human-Technology Interaction group at Eindhoven University of Technology in The Netherlands. Lakens’ work focuses on improving research methods and statistical inferences in the social sciences. He has published more than 100 peer-reviewed articles, including highly cited papers on effect sizes, sequential analyses, equivalence testing, and sample size justification. He won the Ammodo Science award for fundamental research in the Social Sciences in 2023, and the Leamer-Rosenthal Prize for Open Social Science in the category “Leader in Education” in 2017. He created a popular massive open online course and accompanying textbook “Improving Your Statistical Inferences” that he has been used by tens of thousands of people. He is currently funded by an NWO VICI grant for a project on the meaningful interpretation of effect sizes.