You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
每个failure units的metric是怎么选定的? the experienced engineers of an online service system can define the candidate failure units by summarizing the indicative metric groups on different component classes. 可否理解为每个failure units的指标群是人工选择的?
为什么不能用GCN(3.3)
您在论文中提到“ For generalizability, the feature aggregator should be structure-independent. Thus, graph convolutional network (GCN) [29] is unsuitable.”能否详细解释一下为什么要考虑structure-independent
the experienced engineers of an online service system can define the candidate failure units by summarizing the indicative metric groups on different component classes.可否理解为每个failure units的指标群是人工选择的?您在论文中提到“ For generalizability, the feature aggregator should be structure-independent. Thus, graph convolutional network (GCN) [29] is unsuitable.”能否详细解释一下为什么要考虑structure-independent
您在论文中有提到组件、部署关系会经常改变,这意味着FDG需要经常更新,这部分更新使利用专家经验完成吗?
无向图表达相互影响,而非单方面影响(依赖),为什么能反应依赖关系呢?
期待您的解答(╹▽╹)