摘要
近年来许多学者聚焦于研究因果关系的深层原因,根源于人工智能界在深度学习上对模型稳健型的需求,映射出统计学方法在因果问题上面临困境;因果关系在人工智能领域具有不同于哲学语境下的特殊关注点,应该进行跨学科研究界定因果问题的内涵;在哲学语境下从AI领域角度讨论导致因果研究陷于困境的独立同分布问题,可能是近期使得研究取得进展的方向。
In recent years,the reason why many scholars have focused on investigating the deeper causes of causality lies in the needs of the AI community for robust models on deep learning,and it reflects the dilemma faced by statistical methods on causality.As causality has a special focus in the AI domain that is different from the philosophical context,interdisciplinary research should be conducted to define the connotation of causality.Discussing the problem of independent identical distribution from the AI domain in the philosophical context,which has led to the dilemma of causality research,may be the direction of reseach in which progress may be made in the near future.
作者
邱德钧
QIU Dejun(School of Philosophy and Sociology,Lanzhou University,Lanzhou,Gansu,730000)
出处
《自然辩证法通讯》
CSSCI
北大核心
2022年第4期37-43,共7页
Journal of Dialectics of Nature
基金
国家社科基金一般项目“人工智能中关于因果关系的归纳模型研究”(项目编号:20BZX107)。
关键词
因果
稳健性
泛化
独立同分布
Causality
Robustness
Generalization
Independent identical distribution