arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1207.4135cs.AI
收藏

结构化概率模型的事例-因子图

Case-Factor Diagrams for Structured Probabilistic Modeling

TTI at Chicago(芝加哥丰田技术研究所);Massachusetts Institute of Technology(麻省理工学院);University of Pennsylvania(宾夕法尼亚大学)

David A. McAllester, Michael Collins, Fernando Pereira

摘要:

我们引入一个概率形式主义,包括马尔可夫随机场的有界树宽度和概率上下文无关文法。我们的模型基于布尔公式的表示,我们称之为案例因子图(CFD)。CFD类似于二元决策图(BDD),但对于有界树宽度的电路来说是简洁的(不像BDD),并且可以在给定的上下文无关文法下简洁地表示给定字符串上的解析树集合(也不像BDD)。一个概率模型由一个CFD组成,它定义了一组可行的布尔分配和每个布尔变量的权重(或成本)。我们给出了同时计算每个布尔变量的边际的insideoutside算法和求最小代价变量分配的Viterbi算法。这两种算法的运行时间与CFD的大小成比例。

英文摘要:

We introduce a probabilistic formalism subsuming Markov random fields of bounded tree width and probabilistic context free grammars. Our models are based on a representation of Boolean formulas that we call case-factor diagrams (CFDs). CFDs are similar to binary decision diagrams (BDDs) but are concise for circuits of bounded tree width (unlike BDDs) and can concisely represent the set of parse trees over a given string undera given context free grammar (also unlike BDDs). A probabilistic model consists of aCFD defining a feasible set of Boolean assignments and a weight (or cost) for each individual Boolean variable. We give an insideoutside algorithm for simultaneously computing the marginal of each Boolean variable, and a Viterbi algorithm for finding the mininum cost variable assignment. Both algorithms run in time proportional to the size of the CFD.

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。

↑