The input data pipeline is an essential component of each machine learning
(ML) training job. It is responsible for reading massive amounts of training
data, processing batches of samples using complex of transformations, and
loading them onto training nodes at low latency and high throughput. Performant
input data systems are becoming increasingly critical, driven by skyrocketing
data volumes and training throughput demands. Unfortunately, current input data
systems cannot fully leverage key performance optimizations, resulting in
hugely inefficient infrastructures that require significant resources -- or
worse -- underutilize expensive accelerators.
To address these demands, we present cedar, a programming model and framework
that allows users to easily build, optimize, and execute input data pipelines.
cedar presents an easy-to-use programming interface, allowing users to define
input data pipelines using composable operators that support arbitrary ML
frameworks and libraries. Meanwhile, cedar transparently applies a complex and
extensible set of optimization techniques (e.g., offloading, caching,
prefetching, fusion, and reordering). It then orchestrates processing across a
customizable set of local and distributed compute resources in order to
maximize processing performance and efficiency, all without user input. On
average across six diverse input data pipelines, cedar achieves a 2.49x, 1.87x,
2.18x, and 2.74x higher performance compared to tf.data, tf.data service, Ray
Data, and PyTorch's DataLoader, respectively.

cedar 是一个编程模型和框架，允许用户轻松构建、优化和执行输入数据管道，并在六个不同输入数据管道中与 tf.data、tf.data 服务、Ray Data 和 PyTorch 的 DataLoader 相比，平均获得 2.49 倍、1.87 倍、2.18 倍和 2.74 倍更高的性能。

cedar：可组合和优化的机器学习输入数据流程

cedar: Composable and Optimized Machine Learning Input Data Pipelines

IoT as a domain has grown so much in the last few years that it rivals that
of the mobile network environments in terms of data volumes as well as
cybersecurity threats. The confidentiality and privacy of data within IoT
environments have become very important areas of security research within the
last few years. More and more security experts are interested in designing
robust IDS systems to protect IoT environments as a supplement to the more
traditional security methods. Given that IoT devices are resource-constrained
and have a heterogeneous protocol stack, most traditional intrusion detection
approaches don't work well within these schematic boundaries. This has led
security researchers to innovate at the intersection of Machine Learning and
IDS to solve the shortcomings of non-learning based IDS systems in the IoT
ecosystem.
Despite various ML algorithms already having high accuracy with IoT datasets,
we can see a lack of sufficient production grade models. This survey paper
details a comprehensive summary of the latest learning-based approaches used in
IoT intrusion detection systems, and conducts a thorough critical review of
these systems, potential pitfalls in ML pipelines, challenges from an ML
perspective, and discusses future research scope and recommendations.

在物联网领域，由于数据量和网络威胁等方面的增长，数据的机密性和隐私性已经成为安全研究的重要领域。针对物联网环境中的安全性问题，越来越多的安全专家对设计强大的入侵检测系统感兴趣，以作为传统安全方法的补充。本综述论文总结了应用于物联网入侵检测系统的最新基于机器学习的方法，并对这些系统进行了全面的批判性回顾，探讨了机器学习流程中的潜在风险、从机器学习角度面临的挑战，并对未来的研究范围和建议进行了讨论。