Conditional independence encoded in Bayesian networks (BNs) avoids combinatorial explosion on the number of variables. However, BNs are still subject to exponential growth of space and inference time on the number of causes per effect variable in conditional probability tables. A number of space-efficient local models exist that allow efficient encoding of dependency between an effect and its causes, and can also be exploited for improved inference efficiency. We focus on the Nonimpeding Noisy-AND Tree (NIN-AND Tree or NAT) models because of multiple merits. We present a novel framework, trans-causalization of NAT-modeled BNs, by which causal independence embedded in NAT models is exploited for more efficient inference. We show that trans-causalization is exact and yields polynomial space complexity. We demonstrate significant efficiency gain on inference based on lazy propagation and sum–product networks.

中文翻译：

---

跨因果化 NAT 模型贝叶斯网络

在贝叶斯网络 (BN) 中编码的条件独立性避免了变量数量的组合爆炸。然而，在条件概率表中，BN 仍然会随着每个效应变量的原因数量的空间和推理时间呈指数增长。存在许多节省空间的局部模型，这些模型允许对结果与其原因之间的依赖关系进行有效编码，并且还可以用于提高推理效率。由于多重优点，我们专注于无阻碍噪声与树（NIN-AND 树或 NAT）模型。我们提出了一个新颖的框架，NAT 模型 BN 的跨因果化，利用嵌入在 NAT 模型中的因果独立性来进行更有效的推理。我们表明跨因果关系是精确的并且产生多项式空间复杂度。我们证明了基于惰性传播和和积网络的推理效率显着提高。