Header

Machine Learning & Artificial Intelligence — Zurich Center for Market Design

Machine Learning & Artificial Intelligence

Smarter Tools for Better Market Design

Overview

Algorithms in strategic environments

AI and machine learning are reshaping how we build and evaluate market mechanisms. From preference elicitation in complex environments to learning-augmented allocation and strategy-aware algorithms, data-driven tools can boost both performance and usability.

Machine Learning and AI

Research Papers

At the Zurich Center for Market Design, we study how modern AI complements economic theory and mechanism design — aiming for systems that are efficient, fair, and robust to strategic behavior.

Below, we showcase a selection of our recent research at the intersection of market design, artificial intelligence, and machine learning.

LLM-Powered Preference Elicitation in Combinatorial Assignment

Ermis Soumalias, Yanchen Jiang, Kehang Zhu, Michael Curry, Sven Seuken, David C. Parkes

Can large language models simplify preference elicitation in complex allocation tasks? This paper shows how LLM proxies improve efficiency in course assignment while reducing the burden on users.

Click to explore more →

Prices, Bids, Values: One ML-Powered Combinatorial Auction to Rule Them All

Ermis Soumalias, Jakob Heiss, Jakob Weissteiner, Sven Seuken

How can we design more powerful iterative combinatorial auctions? The MLHCA format reduces query complexity and achieves major efficiency gains, cutting welfare loss substantially in experiments.

Click to explore more →

Machine Learning-Powered Course Allocation

Ermis Soumalias, Behnoosh Zamanlooy, Jakob Weissteiner, Sven Seuken

Students often err when reporting course preferences. MLCM uses targeted, learning-guided queries to increase fairness and utility with minimal user effort.

Click to explore more →

Truthful Aggregation of LLMs with an Application to Online Advertising

Ermis Soumalias, Michael J. Curry, Sven Seuken

Platforms need to combine LLM outputs while respecting incentives. MOSAIC is a general-purpose auction mechanism that truthfully aggregates preferences over LLM content and delivers strong welfare and revenue.

Click to explore more →

Learning to Steer Markovian Agents Under Model Uncertainty

Jiawei Huang, Vinzenz Thoma, Zebang Shen, Heinrich Nax, Niao He

How do we design incentives when agents adapt over time and their learning rules are unknown? This work develops theory and RL algorithms to steer multi-agent systems under uncertainty and history dependence.

Click to explore more →