BriefGPT.xyz
Aug, 2020
概率逻辑编程的MAP推断
MAP Inference for Probabilistic Logic Programming
HTML
PDF
Elena Bellodi, Marco Alberti, Fabrizio Riguzzi, Riccardo Zese
TL;DR
本文探讨了在概率逻辑编程中计算最大后验概率和最可能解释的问题，并提出了一种将问题表示为二叉决策图并在其上应用动态规划过程的新算法，与ProbLog在多个合成数据集上的实验结果相比，PITA的性能更佳。
Abstract
In
probabilistic logic programming
(PLP) the most commonly studied inference task is to compute the marginal probability of a query given a program. In this paper, we consider two other important tasks in the PLP setting: the
→