Probabilistic Networks for Automated Reasoning

这是一种基于成就的更新，基于以前在计算实现中的概率推理和定性概率的工作。在因果信息的获取和处理中，需要解决一些重要的问题。这包括(1)通过结合观察和操作从经验数据中学习因果结构，以及(2)通过使用概率和效用的定性近似在不确定性下进行规划。要追求的方法是间接实验、语法推导和反馈机制。预计短期结果将是在医疗诊断、传感器融合以及制造过程自动化管理和控制等应用中改进推理方法。

This is an accomplishment-based renewal based on previous work in probabilistic reasoning and qualitative probabilities within computational implementations. Some important problems are to be addressed in the acquisition and processing of causal information. These include (1) learning of causal structures from empirical data using combinations of observations and manipulations and (2) planning under uncertainty by using qualitative approximations of probabilities and utilities. The approaches to be pursued are indirect experimentation, syntactic derivations , and feedback mechanisms. The short term results are expected to be improved reasoning methods in applications such as medical diagnosis, sensor fusion, and automated management and control of manufacturing processes.