Prior-independent Dynamic Auctions for a Value-maximizing Buyer

Authors: Yuan Deng, Hanrui Zhang

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical Our main result is a novel dynamic mechanism achieving O(T 2/3) regret against a strategic automated bidder, where T is the time horizon. Our dynamic mechanism adopts the explore-and-exploit scheme by first estimating the bidder s value distribution. We then offer a robust version of the single-stage revenue-optimal mechanism to extract the revenue. A cornerstone of our dynamic mechanism in the exploration phase is a novel prior-independent single-stage mechanism that is incentive-compatible for automated bidders. This mechanism forces the bidder to make a trade-off between her future gain from misreporting her values and her immediate loss, leading to a Wasserstein distance bound on the magnitude of misreporting from the automated bidder which is precisely the form of error that our exploitation mechanisms are robust against.
Researcher Affiliation Collaboration Yuan Deng Google Research dengyuan@google.com Hanrui Zhang Duke University hrzhang@cs.duke.edu
Pseudocode Yes Figure 1: A prior-indepentent no-regret mechanism for a value-maximizing buyer.
Open Source Code No The paper is theoretical and does not mention releasing any open-source code for the described methodology.
Open Datasets No The paper does not use a publicly available or open dataset. It defines a theoretical 'value distribution D ([0, 1])' for its analysis.
Dataset Splits No The paper is theoretical and does not involve empirical validation with training, validation, and test splits.
Hardware Specification No The paper describes a theoretical mechanism and does not mention any hardware specifications used for experiments.
Software Dependencies No The paper describes a theoretical mechanism and does not mention any software dependencies with version numbers.
Experiment Setup No The paper describes a theoretical mechanism and does not provide details about an experimental setup, hyperparameters, or training settings.