Risk and Hazard Analysis for Neural Network Systems
Methods and Procedures for the Verification and Validation of Artificial Neural Networks(2006)
摘要
Guidance on risk assessment specific to the special characteristics of neural network software is required, and this is not presently available. General frameworks and taxonomies for software probabilistic risk assessment have been proposed, but they are not specific, nor applicable in some instances, to neural networks. This chapter, while limited in its applicability to every safety- and mission-critical neural network system, points out several possible risk assessment techniques as well as useful places to begin considering identification of risks and hazards.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络