Robust Budget Pacing with a Single Sample

TL;DR

This paper demonstrates that in budget pacing for online advertising, a single sample per distribution suffices to achieve near-optimal regret, significantly reducing data requirements compared to previous methods.

Contribution

The paper introduces a novel approach showing that only one sample per distribution is needed for near-optimal budget pacing, improving over the previous $T ext{log} T$ sample requirement.

Findings

Achieves near-optimal $ ilde O(\sqrt{T})$ regret with one sample per distribution.

Robust to noise in sampling distributions.

Significantly reduces data requirements for budget pacing algorithms.

Abstract

Major Internet advertising platforms offer budget pacing tools as a standard service for advertisers to manage their ad campaigns. Given the inherent non-stationarity in an advertiser's value and also competing advertisers' values over time, a commonly used approach is to learn a target expenditure plan that specifies a target spend as a function of time, and then run a controller that tracks this plan. This raises the question: how many historical samples are required to learn a good expenditure plan? We study this question by considering an advertiser repeatedly participating in $T$ second-price auctions, where the tuple of her value and the highest competing bid is drawn from an unknown time-varying distribution. The advertiser seeks to maximize her total utility subject to her budget constraint. Prior work has shown the sufficiency of $T\log T$ samples per distribution to achieve the optimal $O(\sqrt{T})$-regret. We dramatically improve this state-of-the-art and show that just one sample per distribution is enough to achieve the near-optimal $\tilde O(\sqrt{T})$-regret, while still being robust to noise in the sampling distributions.