贝叶斯推理的认知困境及其消解

发布时间：2018-06-15 03:26

本文选题：贝叶斯推理 + 贝叶斯问题　；参考：《科学技术哲学研究》2017年06期

【摘要】：贝叶斯推理是以贝叶斯定理为基础的一种归纳推理形式,它可以成功避免经典统计推理中的主观因素问题和先验回避问题,并有效消解科学确证、科学接受和归纳逻辑中的诸多悖论和难题,因而在现代归纳逻辑发展史上占据着重要地位,并形成了一种关于归纳推理研究的"贝叶斯主义纲领"。但贝叶斯推理会导致许多与直观概率判断相悖的结论,从而形成诸多的认知困境。这就对贝叶斯推理的有效性形成了挑战。借助认知科学的知识可知:认知困境的产生并不是因为贝叶斯推理本身有问题,而是因为人们在进行直观判断时倾向于采用启发式判断策略。通过采用"使用自然频率来表征概率信息"和"使用无代表性内容的问题"等方式可以有效消解此类认知困境。
[Abstract]:Bayesian reasoning is a form of inductive reasoning based on Bayesian Theorem. It can successfully avoid subjective factors and priori avoidance problems in classical statistical reasoning and effectively eliminate scientific confirmation. Scientific acceptance of many paradoxes and problems in inductive logic, which occupies an important position in the history of modern inductive logic, and forms a "Bayesian creed" for the study of inductive reasoning. However, Bayesian reasoning will lead to many conclusions contrary to intuitive probability judgment, thus forming many cognitive dilemmas. This challenges the effectiveness of Bayesian reasoning. With the help of the knowledge of cognitive science, we know that the cognitive dilemma is not caused by the Bayesian reasoning itself, but because people tend to use heuristic judgment strategies in intuitive judgment. Such cognitive dilemmas can be effectively resolved by using natural frequency to represent probability information and using unrepresentative content.
【作者单位】：西南大学逻辑与智能研究中心;
【基金】：重庆市人文社会科学重点研究基地重点项目(13SKB019) 国家社科基金重大项目(14ZDB016);国家社科基金后期资助项目(15FZX032) 中央高校基本科研业务费专项资金资助(SWU1709602)
【分类号】：O212.8

【相似文献】