Learning and Communications Co-Design for Remote Inference Systems: Feature Length Selection and Transmission Scheduling
Learning and Communications Co-Design for Remote Inference Systems: Feature Length Selection and Transmission Scheduling
复制标题
DOI：
10.1109/jsait.2023.3322620
复制
发表时间：
2023-08
期刊：
影响因子：
--
通讯作者：
Md Kamran Chowdhury Shisher;Bo Ji;I.-Hong Hou;Yin Sun
中科院分区：
文献类型：
--
作者：
Md Kamran Chowdhury Shisher;Bo Ji;I.-Hong Hou;Yin Sun
在本文中，我们考虑一个远程推理系统，其中神经网络用于推断时变目标（例如，机器人运动），基于特征（例如，视频剪辑）从感测节点渐进地接收（例如，照相机）。每个特征都是感官数据的时间序列。推断错误由（i）及时性和（ii）特征的序列长度确定，其中我们使用信息年龄（AoI）作为及时性的度量。虽然较长的特征通常可以提供更好的推断性能，但它通常需要更多的通道资源来发送特征。为了最小化时间平均推理误差，我们研究了一个学习和通信协同设计问题，共同优化特征长度选择和传输调度。当有一个单一的传感器预测器对和一个单一的通道，我们开发了低复杂度的最佳协同设计的情况下，时不变和时变的特征长度。当存在多个传感器-预测器对和多个通道时，协同设计问题变成了PSPACE困难的无休止的多臂多动作强盗问题。对于这种情况，我们设计了一个低复杂度的算法来解决这个问题。跟踪驱动的评估表明，这些协同设计的潜力，以减少高达10000倍的推理错误。
In this paper, we consider a remote inference system, where a neural network is used to infer a time-varying target (e.g., robot movement), based on features (e.g., video clips) that are progressively received from a sensing node (e.g., a camera). Each feature is a temporal sequence of sensory data. The inference error is determined by (i) the timeliness and (ii) the sequence length of the feature, where we use Age of Information (AoI) as a metric for timeliness. While a longer feature can typically provide better inference performance, it often requires more channel resources for sending the feature. To minimize the time-averaged inference error, we study a learning and communication co-design problem that jointly optimizes feature length selection and transmission scheduling. When there is a single sensor-predictor pair and a single channel, we develop low-complexity optimal co-designs for both the cases of time-invariant and time-variant feature length. When there are multiple sensor-predictor pairs and multiple channels, the co-design problem becomes a restless multi-arm multi-action bandit problem that is PSPACE-hard. For this setting, we design a low-complexity algorithm to solve the problem. Trace-driven evaluations demonstrate the potential of these co-designs to reduce inference error by up to 10000 times.