Rational Intelligence Seminar Series: Sebastian Zezulka (Tübingen)
59 views
Skip to first unread message
Krikamol Muandet
Jun 18, 2025, 8:11:30 AM6/18/25
to Machine Learning News
Hi everyone,
We’re thrilled to announce the launch of the Rational Intelligence Seminar Series (RISS), starting June 18, 2025! RISS will bring you cutting-edge talks exploring the crucial connections between AI, decision-making, and causality. Our mission with this series is to deepen our understanding of what makes machine learning systems rational, efficient, and reliable.
Join us to engage in lively discussions in the first session, “Prediction, Potential Outcomes, and Performativity,” delivered by Sebastian Zezulka, a doctoral student at the University of Tübingen.
Abstract
When predictions inform algorithmic policies, they become causal interventions rather than passive forecasts. This performativity entangles pragmatic and epistemic concerns, rendering standard notions of accuracy ambiguous. Common proposals—such as “endogenizing” performative effects or steering outcomes—are misguided. By distinguishing actual from counterfactual predictions, we can separate decision-focused from knowledge-focused issues, enabling models that support both effective choices and informed discourse.
Speaker Bio
Sebastian Zezulka is a doctoral student at the University of Tübingen working on the normative evaluation of algorithmically informed policies, performativity in algorithmic fairness, and evidence-based policymaking. He studied Philosophy & Economics at the University of Bayreuth, Philosophy of the Social Sciences at LSE, and Psychometrics, Econometrics, and Machine Learning at the University of Tübingen.
Logistics
Best regards,
The Rational Intelligence Lab Team
CISPA Helmholtz Center for Information Security
We’re thrilled to announce the launch of the Rational Intelligence Seminar Series (RISS), starting June 18, 2025! RISS will bring you cutting-edge talks exploring the crucial connections between AI, decision-making, and causality. Our mission with this series is to deepen our understanding of what makes machine learning systems rational, efficient, and reliable.
Join us to engage in lively discussions in the first session, “Prediction, Potential Outcomes, and Performativity,” delivered by Sebastian Zezulka, a doctoral student at the University of Tübingen.
Abstract
When predictions inform algorithmic policies, they become causal interventions rather than passive forecasts. This performativity entangles pragmatic and epistemic concerns, rendering standard notions of accuracy ambiguous. Common proposals—such as “endogenizing” performative effects or steering outcomes—are misguided. By distinguishing actual from counterfactual predictions, we can separate decision-focused from knowledge-focused issues, enabling models that support both effective choices and informed discourse.
Speaker Bio
Sebastian Zezulka is a doctoral student at the University of Tübingen working on the normative evaluation of algorithmically informed policies, performativity in algorithmic fairness, and evidence-based policymaking. He studied Philosophy & Economics at the University of Bayreuth, Philosophy of the Social Sciences at LSE, and Psychometrics, Econometrics, and Machine Learning at the University of Tübingen.
Logistics
- Date: June 18, 2025
- Time: 14:30 CET
- Zoom: https://cispa-de.zoom-x.de/j/61708401597
Meeting ID: 617 0840 1597
Best regards,
The Rational Intelligence Lab Team
CISPA Helmholtz Center for Information Security
Reply all
Reply to author
Forward
0 new messages