arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1907.04592cs.AIcs.LO
收藏

可微概率逻辑网络

Differentiable Probabilistic Logic Networks

SingularityNET Foundation(SingularityNET基金会);ITMO University(ITMO大学)

Alexey Potapov, Anatoly Belikov, Vitaly Bogdanov, Alexander Scherbatiy

摘要:

概率逻辑推理是OpenCog等认知架构的核心组件。然而,作为一个综合架构,OpenCog通过混合不同的推理方法促进认知协同。在本文中,我们介绍了一个可微版本的概率逻辑网络,规则操作张量真值的方式,一个链的推理步骤构建一个计算图的张量,接受真值的前提知识库作为输入,并产生真值的结论作为输出。这允许通过结合子符号优化和符号推理的反向传播来学习前提的真值和规则的公式(以具有可训练权重的形式指定)。

英文摘要:

Probabilistic logic reasoning is a central component of such cognitive architectures as OpenCog. However, as an integrative architecture, OpenCog facilitates cognitive synergy via hybridization of different inference methods. In this paper, we introduce a differentiable version of Probabilistic Logic networks, which rules operate over tensor truth values in such a way that a chain of reasoning steps constructs a computation graph over tensors that accepts truth values of premises from the knowledge base as input and produces truth values of conclusions as output. This allows for both learning truth values of premises and formulas for rules (specified in a form with trainable weights) by backpropagation combining subsymbolic optimization and symbolic reasoning.

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。

↑