Causality for Ethics and Society (24 - 25 July 2023)
Idea & Motivation
Causality plays a fundamental role in understanding many key ethical notions, such as responsibility, interpretability, fairness, harm, and related concepts. Causality is also fundamental to understanding broader social challenges, such as discrimination, inequality, the impact of artificial intelligence on society, and others. The rise of causal modeling methods has opened up entirely novel ways of thinking about the role that causality plays in addressing all of these issues, with the potential to both benefit from and contribute to a wide range of disciplines.
Causal models are already being actively employed in addressing important social challenges. In the sociological and legal literature on discrimination, causal methodology has been central both to conceptual discussions of whether demographic variables such as race, gender, and age can be causes, and in addressing thorny methodological disputes about how to detect discrimination using statistical data. Within philosophy and artificial intelligence, causal models have provided a basis for more rigorously theorizing about causal explanations, moral responsibility, and harm. Finally, causal models have contributed to discussions of algorithmic fairness by combining traditional statistical approaches to discrimination with causal approaches that focus on the use of proxies, path-specific effects, and counterfactuals.
This workshop will bring together participants from a range of academic fields in order to present and discuss the most recent developments on employing formal causal reasoning in ethics and in social contexts. Possible topics include, but are not limited to the following topics.
• Algorithmic fairness, harm, or bias, through a causal lens
• Causal models and moral responsibility
• The causal analysis of discrimination
• Applications to AI and deep learning
• The role of values in causal modeling
In addition to these possibilities, we will consider any submission employing causal models in addressing an ethical or socially relevant problem. We welcome contributions debating the applicability of causal concepts within a particular domain as well as proposals for modifying existing causal frameworks for addressing novel problems. As the workshop will include speakers from a wide range of academic fields, we encourage contributions that address problems arising in interdisciplinary contexts.
Confirmed Speakers
- Elias Bareinboim (Computer Science, Columbia University)
- Joseph Halpern (Computer Science, Cornell University)
- Lily Hu (Philosophy, Yale University)
- Niki Kilbertus (Computer Science, TU Munich)
- Kasper Lippert-Rasmussen (Political Science, Aarhus University)
Registration
Location
LMU München
Prof.-Huber-Pl. 2 (W) - LEHRTURM
Room W401
80539 München
Program
Day 1 (23 July 2023)
Time | Event |
---|---|
09:00 - 09:30 | Registration |
09:30 - 09:45 | Introduction |
09:45 - 11:00 | Keynote 1: Lily Hu |
11:00 - 11:15 | Coffee Break |
11:15 - 12:45 | Session 1: Lennart Ackermans, Dean McHugh, Ludwig Bothmann |
12:45 - 14:30 | Lunch Break - lunch not provided |
14:30 - 16:00 | Session 2: Nicol'o Cangiotti, Charles Wan, Michele Loi |
16:00 - 16:15 | Coffee Break |
16:15 - 16:45 | Session 3: Tzvetan Moev |
16:45 - 18:00 | Keynote 2: Elias Bareinboim |
18:30 - 20:00 | Reception at Taverna Kalypso (15 min walk) |
Day 2 (24 July 2023)
Time | Event |
---|---|
09:45 - 11:00 | Keynote 3: Kasper Lippert-Rasmussen |
11:00 - 11:15 | Coffee Break |
11:15 - 12:45 | Session 4: Francesco Nappo, David Kinney, Sebastian Zezulka |
12:45 - 14:30 | Lunch Break |
14:30 - 15:45 | Keynote 4: Niki Kilbertus |
15:45 - 16:15 | Session 5: Yuval Abrams |
16:15 - 16:30 | Coffee Break |
16:30 - 17:45 | Keynote 5: Joseph Halpern |
Abstracts
Keynote 1: TBD, Lily Hu (Yale University)
TBAtop
Session 1:
Defending the Perception Approach to Measuring Discrimination, Lennart Ackermans (Erasmus University Rotterdam)
A New Test for Discrimination: the Existential But-For Test, Dean McHugh (University of Amsterdam)
Causal Fair Machine Learning via Warping Real World Data to a Fictitious, Normatively Desired World, Ludwig Bothmann (LMU Munich)
Session 2:
Classification Parity, Causal Equal Protection and Algorithmic Fairness, Nicol Cangiotti (Politecnico di Milano), Marcello Di Bello (Arizona State University), and Michele Loi (Politecnico di Milano)
How Differential Robustness Creates Disparate Impact, Charles Wan (Erasmus University
Rotterdam)
Algorithmic Unfairness and Group Level Causation, Michele Loi (Politecnico di Milano)top
Session 3:
Why do methodologists disagree about how to do causal inference in social science?,
Tzvetan Moev (Duke University)top
Keynote 2: TBD, Elias Bareinboim (Columbia University)
TBAtop
Keynote 3: Algorithmic and Non-Algorithmic Fairness: Should We Revise our View of the
Latter on our Account of Our View of the Former?, Kasper Lippert-Rasmussen (Aarhus University)
Session 4:
How I Would Have Been Differently Treated. Discrimination Through the Lens of Counterfactual Fairness, Francesco Nappo (Politecnico di Milano), Michele Loi (Politecnico di Milano), and Eleonora Vigan'o (University of Zurich)
Identifying Fair and Good Predictors, David Kinney (Princeton University)
Fairness after Intervention, Sebastian Zezulka (University of Tübingen)top
Keynote 4: TBD, Niki Kilbertus (TU Munich)
TBAtop
Session 5: Value Free Causation, Yuval Abrams (New York University)
Keynote 5: A Causal Analysis of Harm, Joseph Halpern (Cornell University)
TBA
Organizers
- Sander Beckers (University of Amsterdam)
- Naftali Weinberger (MCMP/LMU Munich)
General questions about the conference can be sent to Causalityforsociety@gmail.com.
Acknowledgement
The conference is supported by the Alexander von Humboldt - Foundation.