arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2410.08710cs.LGstat.ML
收藏

偏好归一化流

Preferential Normalizing Flows

University of Helsinki(赫尔辛基大学)

Petrus Mikkola, Luigi Acerbi, Arto Klami

AI总结:

提出仅基于偏好问题将专家信念密度作为归一化流获取的方法,引入泛函先验解决概率质量坍缩或发散问题,实证表明可作为函数空间最大后验估计推断出密度。

摘要:

通过含噪判断从专家处获取高维概率分布是极具挑战性的,但这对先验获取和奖励建模等许多应用非常有用。我们提出了一种仅基于比较或排序备选方案等偏好问题,将专家信念密度作为归一化流获取的方法。这原则上允许获取任意灵活的密度,但流估计易受概率质量坍缩或发散的影响,导致实践中困难重重。我们通过引入一种新的流泛函先验来解决此问题,该先验由决策理论论证驱动,并实证表明信念密度可作为函数空间最大后验估计被推断出来。我们通过获取模拟专家的多变量信念密度来演示该方法,包括通用大语言模型在真实数据集上的先验信念。

英文摘要:

Eliciting a high-dimensional probability distribution from an expert via noisy judgments is notoriously challenging, yet useful for many applications, such as prior elicitation and reward modeling. We introduce a method for eliciting the expert's belief density as a normalizing flow based solely on preferential questions such as comparing or ranking alternatives. This allows eliciting in principle arbitrarily flexible densities, but flow estimation is susceptible to the challenge of collapsing or diverging probability mass that makes it difficult in practice. We tackle this problem by introducing a novel functional prior for the flow, motivated by a decision-theoretic argument, and show empirically that the belief density can be inferred as the function-space maximum a posteriori estimate. We demonstrate our method by eliciting multivariate belief densities of simulated experts, including the prior belief of a general-purpose large language model over a real-world dataset.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑