Celestine Mendler-Dünner

Selected Talks and Highlights

• 09/2025
  Speaking at the Swiss CLOCK Summit in Engelberg
• 06/2025
  Snap ML reached 1 million downloads on pypi!
• 04/2025
  "Performative Prediction: Past and Future'' accepted to Statistical Science
• 12/2024
  NeurIPS oral -- join us for Ricardo's talk on LLMs and surveys in Vancouver
• 12/2024
  Speaking at the MPG AI & Research event in Berlin
• 11/2024
  Speaking at the ETH Seminar on Law & Economics and the ML meets Law Workshop in Tübingen
• 09/2024
  Keynote talk at the Cornell Thought Summit on the Future of Survey Science
• 07/2024
  Invited tutorial on Performative Prediction at UAI with Tijana
• 06/2024
  Speaking at the ELLIS Institute Opening
• 12/2023
  Invited talk at the NeurIPS Workshop on Algorithmic Fairness
• 11/2023
  Serving as Program Chair for EAAMO'23
• 05/2023
  Speaking at the Oxford Workshop for Social Foundations of Statistics and ML
• 02/2023
  Speaking at the TILOS-OPTML++ Seminar at MIT

Affiliations and Awards

Advising

Featured Research Projects

LLMs and Surveys Performative Prediction Power in Digital Economies Algorithmic Collective Action Scalable Learning Algorithms

LLMs and Surveys

As large language models increase in capability, researchers have started to conduct surveys of all kinds on these models---research questions rage from learning about biases and alignment of LLMs, to extracting information about human subpopulations and using them as a data source for social science research. In our work we carefully examine the validity of surveys as a method to extract population statistics from LLMs. We explain pitfalls, develop tools for the systematic testing of the capabilities of LLMs to emulate human outcomes, and design statistically rigorous methods to integrate them into the research pipeline.

Evaluating LLMs as risk scores

A Python library to systematically translate the ACS survey data into natural text prompts to benchmark LLM outputs against US population statistics. Individual prediction tasks are non-realizable and provide insights into model's ability to express natural uncertainty in human outcomes.

Performative Prediction

Predictions when deployed in societal systems can trigger actions and reactions of individuals. Thereby they can change the way the broader system behaves -- a dynamic effect that traditional machine learning fails to account for. To formalize and reason about performativity in the context of machine learning, we have developed the framework of performative prediction. It extends the classical risk minimization framework in that it allows the data distribution to depend on the deployed model. This inherently dynamic viewpoint leads to new solution concepts and optimization challenges, and it brings forth interesting connections to concepts from causality, game theory and control.

Performative Prediction

• Performative Prediction. ICML 2020.
• Survey paper. Statistical Science (forthcoming).
• Tutorial. UAI 2024.

Power in Digital Economies

Algorithmic predictions play an increasingly important role in digital economies as they mediate services, platforms, and markets at societal scale. The fact that such services can steer consumption and behavior is a central concern in various policy questions. I am exploring how the concept of performativity can help quantify economic power in digital economies and the potential for consumer harm to support digital market investigations. Intuitively, the more powerful a firm the more performative their predictions, a causal effect strength we can measure from observational and experimental data.

A Chrome extension that measure how content arrangements impact user click behavior through randomized experiments.

• Project website at https://powermeter.is.tue.mpg.de/
• Experimental insights published at NeurIPS 2024
• Montreal AI Ethics Institute blog post

Algorithmic Collective Action

How can participants steer AI systems toward a common good? We explore data as a lever for participants to impact learning outcomes from below. Research questions encompass the design of goals and strategies, their effectiveness in different learning settings, levers in AI systems, challenges of coordination, as well as connection to labor markets, power dynamics, game theory and collective action in political economy.

Algorithmic collective action

• Theoretical framework. ICML 2022.
• Application to sequential recommender systems. NeurIPS 2024.
• A case study on ride hailing platforms ACM CHI 2025.
• A collection of documented use-cases. Github.

Scalable Learning Algorithms

When building societal-scale machine learning systems, training time and resource constraints can be a critical bottleneck for dynamic system optimization. Efficient training algorithms that are aware of distributed architectures, interconnect topologies, memory constraints and accelerator units form an important bilding block towards using available resources most efficiently. As part of my PhD research, we demonstrated that system-aware algroithms can lead to several orders of magnitude reduction in training time compared to standard system-agnostic methods. Today, the innovations of my PhD research form the backbone of the IBM Snap ML library that has been integrated with several of IBM's core AI products.

IBM Snap ML is a library that provides resource efficient and fast training of popular machine learning models on modern computing systems.
>900k downloads on PyPi

• Project Website: https://www.zurich.ibm.com/snapml/
• Publication: NeurIPS 2018

Resources, Media and Recordings

• Folktexts datasets available on HuggingFace and Github
• UAI tutorial recording: An Introduction to Performative Prediction
• Powermeter is available in the Chrome store
• Invited talk NeurIPS workshop on Algorithmic Fairness Through the Lense of Time.
• Invited talk Oxford Workshop on Social Foundations of Statistics and Machine Learning
• Invited talk MIT Optimization Seminar
• Snap ML in the press: Forbes, The Register and EE Times writing about our research
• Snap ML has been integrated with the AI Toolkit for IBM Z and Watson ML.
• Snap ML available through PyPi: > pip install snapml (see Medium article)
• General audience interview for the Cyber Valley Podcast Series, see below.