通过对比反事实动作，识别真实成功动作，解决推荐系统中的混淆问题。
Offline Contextual Bandit with Counterfactual Sample Identification
- 用反事实动作对比，判断哪个动作导致成功结果
- 在合成数据和真实场景中均优于传统直接奖励模型
- 适合有历史日志数据的推荐与广告系统优化
在生产系统中，上下文贝尔特方法通常依赖于以动作和上下文为输入的直接奖励模型。然而，这些模型容易受到混淆影响，难以区分动作与上下文各自的作用。我们提出「反事实样本识别」，将问题重新建模：不预测奖励，而是学习识别在相同上下文中，哪个动作带来了成功（二元）结果，通过与来自记录策略的反事实动作进行比较。该方法理论基础扎实，在合成实验和真实部署中均持续优于直接模型。
原文摘要 · Abstract (English)
In production systems, contextual bandit approaches often rely on direct reward models that take both action and context as input. However, these models can suffer from confounding, making it difficult to isolate the effect of the action from that of the context. We present \emph{Counterfactual Sample Identification}, a new approach that re-frames the problem: rather than predicting reward, it learns to recognize which action led to a successful (binary) outcome by comparing it to a counterfactual action sampled from the logging policy under the same context. The method is theoretically grounded and consistently outperforms direct models in both synthetic experiments and real-world deployments.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。