Online Ad Procurement in Non-stationary Autobidding Worlds
Not on the map. Why?
Abstract (arXiv)
Today's online advertisers procure digital ad impressions through interacting with autobidding platforms: advertisers convey high level procurement goals via setting levers such as budget, target return-on-investment, max cost per click, etc.. Then ads platforms subsequently procure impressions on advertisers' behalf, and report final procurement conversions (e.g. click) to advertisers. In practice, advertisers may receive minimal information on platforms' procurement details, and procurement outcomes are subject to non-stationary factors like seasonal patterns, occasional system corruptions, and market trends which make it difficult for advertisers to optimize lever decisions effectively. Motivated by this, we present an online learning framework that helps advertisers dynamically optimize ad platform lever decisions while subject to general long-term constraints in a realistic bandit feedback environment with non-stationary procurement outcomes. In particular, we introduce a primal-dual algorithm for online decision making with multi-dimension decision variables, bandit feedback and long-term uncertain constraints. We show that our algorithm achieves low regret in many worlds when procurement outcomes are generated through procedures that are stochastic, adversarial, adversarially corrupted, periodic, and ergodic, respectively, without having to know which procedure is the ground truth. Finally, we emphasize that our proposed algorithm and theoretical results extend beyond the applications of online advertising.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 24 of its 45 references that are in the map. First posted on arXiv 10 Jul 2023.
Builds on
24 of 45 references.
References · 24
- 2013Bandits with Knapsackscited by 163
- 2019Linear Stochastic Bandits Under Safety Constraintscited by 90
- 2015Adaptive Algorithms for Online Convex Optimization with Long-term Constraintscited by 65
- 2020The Best of Many Worlds: Dual Mirror Descent for Online Allocation Problemscited by 58
- 2018Online convex optimization for cumulative constraintscited by 47
- 2017Selling to a No-Regret Buyercited by 43
- 2019Bandit Convex Optimization in Non-stationary Environmentscited by 41
- 2022A Unifying Framework for Online Optimization with Long-Term Constraintscited by 41
- 2015Online learning in repeated auctionscited by 39
- 2020Optimal No-regret Learning in Repeated First-price Auctionscited by 33
- 2020Learning to Bid Optimally and Efficiently in Adversarial First-price Auctionscited by 33
- 2021Robust Auction Design in the Auto-bidding Worldcited by 21
12 more
- 2016An optimal algorithm for bandit convex optimizationcited by 20
- 2021Towards Efficient Auctions in an Auto-bidding Worldcited by 19
- 2018Projection-Free Bandit Convex Optimizationcited by 17
- 2022Online Bidding Algorithms for Return-on-Spend Constrained Advertiserscited by 17
- 2023Multi-channel Autobidding with Budget and ROI Constraintscited by 14
- 2019Incentive-aware Contextual Pricing with Non-parametric Market Noisecited by 13
- 2021Safe Online Convex Optimization with Unknown Linear Safety Constraintscited by 13
- 2020Learning Product Rankings Robust to Fake Userscited by 12
- 2022Auction Design in an Auto-bidding Setting: Randomization Improves Efficiency Beyond VCGcited by 6
- 2022Efficiency of the First-Price Auction in the Autobidding Worldcited by 6
- 2022Fair Assortment Planningcited by 5
- 2020No-regret Learning in Price Competitions under Consumer Reference Effectscited by 3
Built on by
Citing papers per year: 2026: 1.
1 paper.