arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2204.04705cs.LGcs.AIcs.DC
收藏

SplitNets:为头戴式系统上的高效分布式计算设计神经架构

SplitNets: Designing Neural Architectures for Efficient Distributed Computing on Head-Mounted Systems

Harvard University(哈佛大学);Meta Reality Labs(Meta现实实验室);ETH Zurich(苏黎世联邦理工学院)

Xin Dong, Barbara De Salvo, Meng Li, Chiao Liu, Zhongnan Qu, H. T. Kung, Ziyun Li

摘要:

我们设计了深度神经网络(DNN)和相应的网络分裂,将DNN的工作负载分配给摄像头传感器和头戴式设备上的集中式聚合器,以满足给定硬件资源约束下的推理精度和延迟方面的系统性能目标。为了实现计算,通信和性能之间的最佳平衡,引入了分裂感知神经架构搜索框架SplitNets,以同时进行模型设计,分裂和通信简化。我们进一步扩展的框架,多视图系统学习融合输入从多个摄像头传感器具有最佳的性能和系统效率。我们验证了SplitNets在ImageNet上的单视图系统以及在3D分类上的多视图系统,并表明SplitNets框架与现有方法相比实现了最先进的(SOTA)性能和系统延迟。

英文摘要:

We design deep neural networks (DNNs) and corresponding networks' splittings to distribute DNNs' workload to camera sensors and a centralized aggregator on head mounted devices to meet system performance targets in inference accuracy and latency under the given hardware resource constraints. To achieve an optimal balance among computation, communication, and performance, a split-aware neural architecture search framework, SplitNets, is introduced to conduct model designing, splitting, and communication reduction simultaneously. We further extend the framework to multi-view systems for learning to fuse inputs from multiple camera sensors with optimal performance and systemic efficiency. We validate SplitNets for single-view system on ImageNet as well as multi-view system on 3D classification, and show that the SplitNets framework achieves state-of-the-art (SOTA) performance and system latency compared with existing approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑