On Removing Algorithmic Priority Inversion from Mission-critical Machine Inference Pipelines
On Removing Algorithmic Priority Inversion from Mission-critical Machine Inference Pipelines
复制标题
DOI：
10.1109/rtss49844.2020.00037
复制
发表时间：
2020-12
期刊：
影响因子：
--
通讯作者：
Shengzhong Liu;Shuochao Yao;Xinzhe Fu;Rohan Tabish;Simon Yu;Ayoosh Bansal;H. Yun;L. Sha;T. Abdelzaher
中科院分区：
文献类型：
--
作者：
Shengzhong Liu;Shuochao Yao;Xinzhe Fu;Rohan Tabish;Simon Yu;Ayoosh Bansal;H. Yun;L. Sha;T. Abdelzaher
本文讨论了现代基于神经网络的网络物理应用中使用的关键任务机器推理管道中的算法优先级反转，并开发了一种调度解决方案来减轻其影响。一般来说，当较低优先级的计算与较高优先级的计算一起或先于较高优先级的计算一起执行时，优先级反转会发生在实时系统中。1在当前的机器智能软件中，显着的优先级反转发生在从感知到决策的路径上，其中底层神经网络算法的执行不区分关键数据和次关键数据。我们描述了一个调度框架来解决这个问题，并证明它提高了系统对关键输入做出反应的能力，同时降低了平台成本。
The paper discusses algorithmic priority inversion in mission-critical machine inference pipelines used in modern neural-network-based cyber-physical applications, and develops a scheduling solution to mitigate its effect. In general, priority inversion occurs in real-time systems when computations that are of lower priority are performed together with or ahead of those that are of higher priority.1 In current machine intelligence software, significant priority inversion occurs on the path from perception to decision-making, where the execution of underlying neural network algorithms does not differentiate between critical and less critical data. We describe a scheduling framework to resolve this problem, and demonstrate that it improves the system’s ability to react to critical inputs, while at the same time reducing platform cost.