We address the problem of training conversion prediction models in advertising domains under privacy constraints, where direct links between ad clicks and conversions are unavailable. Motivated by privacy-preserving browser APIs and the deprecation of third-party cookies, we study a setting where the learner observes a sequence of clicks and a sequence of conversions, but can only link a conversion to a set of candidate clicks (an attribution set) rather than a unique source. We formalize this as learning from attribution sets generated by an oblivious adversary equipped with a prior distribution over the candidates. Despite the lack of explicit labels, we construct an unbiased estimator of the population loss from these coarse signals via a novel approach. Leveraging this estimator, we show that Empirical Risk Minimization achieves generalization guarantees that scale with the informativeness of the prior and is also robust against estimation errors in the prior, despite complex dependencies among attribution sets. Simple empirical evaluations on standard datasets suggest our unbiased approach significantly outperforms common industry heuristics, particularly in regimes where attribution sets are large or overlapping.

翻译：本文研究在隐私约束下广告领域转化预测模型的训练问题，其中广告点击与转化之间的直接关联不可获取。受隐私保护浏览器API与第三方Cookie弃用的驱动，我们研究一种学习场景：学习者能观测到点击序列与转化序列，但仅能将转化关联至候选点击集合（即归因集合），而非唯一来源。我们将此形式化为从由具备候选先验分布的遗忘对手生成的归因集合中进行学习。尽管缺乏显式标签，我们通过一种新颖方法从这些粗粒度信号中构建出总体损失的无偏估计量。利用该估计量，我们证明经验风险最小化能够获得泛化保证，其效果随先验信息量而扩展，且即使归因集合间存在复杂依赖关系，该方法对先验估计误差仍具鲁棒性。在标准数据集上的简易实证评估表明，我们的无偏方法显著优于行业常见启发式方法，尤其在归因集合规模较大或存在重叠的场景中表现突出。