首页>
外国专利>
Learning methods and devices that provide functional safety by alerting drivers of potential risk situations using explainable artificial intelligence that validates the detection process of autonomous driving networks, and the use of such learning devices. Sting method and testing apparatus {LEARNING mETHOD aND LEARNING dEVICE FOR PROVIDING FUNCTIONAL SAFETY BY WARNING DRIVER ABOUT POTENTIAL DANGEROUS SITUATION BY USING EXPLAINABLE AI WHICH VERIFIES DETECTION PROCESSES OF AUTONOMOUS DRIVING NETWORK, aND tESTING mETHOD aND tESTING dEVICE USING THE SAME}
Learning methods and devices that provide functional safety by alerting drivers of potential risk situations using explainable artificial intelligence that validates the detection process of autonomous driving networks, and the use of such learning devices. Sting method and testing apparatus {LEARNING mETHOD aND LEARNING dEVICE FOR PROVIDING FUNCTIONAL SAFETY BY WARNING DRIVER ABOUT POTENTIAL DANGEROUS SITUATION BY USING EXPLAINABLE AI WHICH VERIFIES DETECTION PROCESSES OF AUTONOMOUS DRIVING NETWORK, aND tESTING mETHOD aND tESTING dEVICE USING THE SAME}
PROBLEM TO BE SOLVED: To provide a learning device for verifying a detection process of a neural network for autonomous driving and providing functional safety by warning a driver of a potentially dangerous situation by using an explainable artificial intelligence. A verification learning device, when acquiring at least one verification training image, has an attribute extraction module and applies an extraction operation to the verification training image to extract attribute information for characteristics of the verification training image. To generate a quality vector. The learning device for verification generates the predicted safety information by applying one or more first neural network operations to the quality vector with the verification neural network, and generates the loss with the loss module. The parameters included in the verification neural network are learned by performing back propagation using. [Selection diagram] Figure 2
展开▼