arXiv cs.AI· Otto Nyberg, Davide Tugnoli, Fausto Carcassi, Giovanni Cin\`a·· 13 小时前AI 评分34
2-Step Agent：贝叶斯决策者如何从 AI 决策支持中学习
2-Step Agent: How a Bayesian Decision Maker Learns from AI-Decision Support
AI 导读
研究者提出 2-Step Agent 计算框架，刻画 ML 模型预测如何同时成为关于训练数据与因果机制的信息来源，并影响贝叶斯智能体的信念、因果效应估计与下游决策。该工作在多元线性-高斯设定下给出可求解的贝叶斯推断方案，实验识别出 ML 决策支持优于无辅助决策的条件。研究还表明，即便预测模型设定正确、智能体完全理性，用户单一的先验信念偏差就足以让 ML-DS 的下游结果劣于无决策支持。
正文 · 原文
Abstract:Prognostic predictions from ML models support human decision making in several fields, including high-stakes ones such as healthcare and the judiciary. Yet, we still lack a clear understanding of how decision makers learn from ML-based decision support (ML-DS). In this paper, we introduce a general computational framework, the 2-Step Agent, to capture how a prediction from an ML model can also be used as a source of useful information about the training data and the causal mechanism that generated it. Our framework models (i) how a prediction from an ML model affects the beliefs of a rational Bayesian agent, and (ii) how this change in beliefs affects the estimation of the causal effect, the downstream decision for a new observation, and the subsequent outcome. In addition to the framework itself, we make three contributions. First, for the multivariate linear-Gaussian setting, we derive a tractable solution for the Bayesian inference problem introduced, i.e. one in which the agent infers information about a population from an ML prediction. Second, we experimentally identify conditions under which ML-DS is beneficial compared to unassisted decision making. Third, we show that a single misaligned prior belief of the user can be sufficient for ML-DS to lead to worse downstream outcomes compared to no decision support even when the prognostic model is well-specified and the agent is perfectly rational. Hence, even under ideal conditions, ML-DS can do more harm than good.
| Comments: | Accepted at NeurIPS 2026. Camera-ready version. 39 pages, 8 figures |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2602.21889 [cs.AI] |
| (or arXiv:2602.21889v4 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2602.21889 arXiv-issued DOI via DataCite |
Submission history
From: Davide Tugnoli [view email]
[v1]
Wed, 25 Feb 2026 13:11:12 UTC (452 KB)
[v2]
Thu, 4 Jun 2026 13:00:30 UTC (2,040 KB)
[v3]
Fri, 5 Jun 2026 18:14:07 UTC (2,040 KB)
[v4]
Wed, 21 Oct 2026 08:13:14 UTC (2,236 KB)
来源：arXiv cs.AI · arxiv.org