[网络领域顶会追踪系列#9][NSDI'22]Ekay：边缘计算服务器上视频分析模型的持续学习

本系列为个人论文阅读笔记，大多数应该是对顶会摘要的个人菜鸟翻译，旨在让感兴趣的朋友能快速浏览、获取信息。如有错误或心得，欢迎在评论区交流指正！

Ekya: Continuous Learning of Video Analytics Models on Edge Compute Servers

背景

视频分析应用使用边缘计算服务器来处理视频。用于推断的压缩模型被部署在边缘服务器上，经历着数据漂移，即实时视频流数据与训练数据发生了偏移。持续学习通过周期性的在新数据上重新训练模型来处理数据漂移问题。

挑战

要能同时支持边缘服务器上的推断和重训练任务，着就需要在重训练模型的准确性和推断准确性之间的基本权衡间做出探索。

工作

提出Ekya，能够在多个模型间的这种权衡中做出平衡，使用了一个微型测绘器（micro-profiler）来识别出最需要重训练的模型。

实验

Ekya的准确度增益比baseline调度器高出29%；并且在实现相同准确度的情况下，baseline需要比Ekya超出4倍的GPU资源。

原文摘要

Video analytics applications use edge compute servers for processing videos. Compressed models that are deployed on the edge servers for inference suffer from data drift where the live video data diverges from the training data. Continuous learning handles data drift by periodically retraining the models on new data. Our work addresses the challenge of jointly supporting inference and retraining tasks on edge servers, which requires navigating the fundamental tradeoff between the retrained model’s accuracy and the inference accuracy. Our solution Ekya balances this tradeoff across multiple models and uses a micro-profiler to identify the models most in need of retraining. Ekya’s accuracy gain compared to a baseline scheduler is 29% higher, and the baseline requires 4× more GPU resources to achieve the same accuracy as Ekya.