arXiv：cs.AI· Ilan Lobel, Humberto Moreira, Omar Mouchtaki·· 5 小时前
当 ML 预测随机出错：区分幻觉与噪声的贝叶斯拍卖设计
Randomly Wrong Signals: Bayesian Auction Design with ML Predictions
AI 导读
研究者提出一种贝叶斯拍卖模型，将 ML 对买家支付意愿的"幻觉式"错误预测与普通测量噪声区分开来。在信息性预测完全准确的基准下，该模型刻画了收益最大化拍卖的形态，并说明单一买家场景中卖家何时应采纳预测、何时应忽略预测或设置价格上限来对冲其失效。数值实验显示，据此设计的简单拍卖可达到接近最优的收益。
正文
Abstract:Machine learning can help sellers predict buyers' willingness to pay, but its predictions may be plausible and yet entirely unrelated to a buyer's value. We introduce a Bayesian auction model that distinguishes such hallucinations from ordinary measurement noise. We characterize revenue-maximizing auctions with publicly disclosed predictions in the benchmark where informative predictions are exact. For a single buyer, this characterization explains when the seller should follow a prediction and when to hedge against its failure by ignoring it or capping the price. These insights guide the design of simple auctions that account for both hallucinations and measurement noise. Numerical experiments show that the resulting auctions achieve near-optimal revenue.
| Subjects: | Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2502.08792 [cs.GT] |
| (or arXiv:2502.08792v4 [cs.GT] for this version) | |
| https://doi.org/10.48550/arXiv.2502.08792 arXiv-issued DOI via DataCite |
Submission history
From: Omar Mouchtaki [view email]
[v1]
Wed, 12 Feb 2025 21:08:28 UTC (135 KB)
[v2]
Mon, 13 Oct 2025 13:49:32 UTC (138 KB)
[v3]
Tue, 27 Jan 2026 19:48:13 UTC (181 KB)
[v4]
Sat, 10 Oct 2026 01:25:10 UTC (206 KB)
来源：arXiv：cs.AI · arxiv.org