Advances in Auction Mechanisms and Online Learning

The field of auction mechanisms and online learning is rapidly evolving, with a focus on developing innovative strategies for heterogeneous auction channels and improving the efficiency of bidding systems. Recent research has explored the use of marginal cost alignment strategies and decentralized online learning algorithms to achieve sublinear regret bounds. Additionally, there has been a significant improvement in the development of online bidding algorithms for repeated first-price auctions with ROI constraints, achieving near-optimal regret bounds. The use of preference-based learning and Vickrey-Clarke-Groves payments has also been proposed to incentivize truthful reporting in resource allocation problems. Noteworthy papers include: HOB: A Holistically Optimized Bidding Strategy, which introduces a marginal cost alignment strategy that provably secures bidding efficiency across heterogeneous auction mechanisms. Decentralized Parameter-Free Online Learning, which proposes the first parameter-free decentralized online learning algorithms with network regret guarantees. No-Regret Online Autobidding Algorithms in First-price Auctions, which develops online bidding algorithms for repeated first-price auctions with ROI constraints, benchmarking against the optimal randomized strategy in hindsight.

Sources

HOB: A Holistically Optimized Bidding Strategy under Heterogeneous Auction Mechanisms with Organic Traffic

Decentralized Parameter-Free Online Learning

No-Regret Online Autobidding Algorithms in First-price Auctions

Eliciting Truthful Feedback for Preference-Based Learning via the VCG Mechanism

Autobidding Arena: unified evaluation of the classical and RL-based autobidding algorithms

Comparing Uniform Price and Discriminatory Multi-Unit Auctions through Regret Minimization

Built with on top of