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

Ermis Soumalias, Jakob Heiss, Jakob Weissteiner, Sven Seuken

International Conference on Machine Learning 2025 · Oral

This talk introduces a groundbreaking hybrid auction mechanism that masterfully integrates two distinct types of queries—**demand queries** and **value queries**—powered by machine learning. Presented by Ermis Soumalias, Jakob Heiss, Jakob Weissteiner, and Sven Seuken, the research addresses the long-standing challenge of maximizing **social welfare** in complex **combinatorial auctions** where bidders' preferences exhibit intricate relationships like substitutes and complements. The core innovation lies in a phased approach that strategically leverages the complementary strengths of demand and value queries, each enhanced by ML models, to explore and then exploit bidder preferences efficiently.

AI review

PBV-MLCCA is a competent and practically motivated hybrid auction mechanism that combines demand and value queries under ML guidance. The work contains one genuinely interesting theoretical result — the 55% efficiency ceiling for demand-query-only mechanisms — and a clever engineering contribution in the Bridge Bid. The empirical results are strong on the standard test suite and the welfare numbers are compelling in applied terms. What holds this back from a higher rating is the article's consistent conflation of heuristic design choices with principled theoretical derivations…