Neural Architecture Search (NAS) is a popular tool for automatically generating Neural Network (NN) architectures. In early NAS works, these tools typically optimized NN architectures for a single metric, such as accuracy. However, in the case of resource constrained Machine Learning, one single metric is not enough to evaluate a NN architecture. For example, a NN model achieving a high accuracy is not useful if it does not fit inside the flash memory of a given system. Therefore, recent works on NAS for resource constrained systems have investigated various approaches to optimize for multiple metrics. In this paper, we propose that, on top of these approaches, it could be beneficial for NAS optimization of resource constrained systems to also consider input data granularity. We name such a system "Data Aware NAS", and we provide experimental evidence of its benefits by comparing it to traditional NAS.

翻译：神经架构搜索（NAS）是一种自动生成神经网络（NN）架构的流行工具。在早期的NAS研究中，这些工具通常针对单一指标（如准确率）优化NN架构。然而，在资源受限的机器学习场景下，单一指标不足以全面评估NN架构的性能。例如，一个高准确率的NN模型若无法适配特定系统的闪存容量，则实际上并无实用价值。因此，近年来面向资源受限系统的NAS研究探索了多种多目标优化方法。本文提出，在上述方法的基础上，将输入数据粒度纳入考量将有助于提升资源受限系统的NAS优化效果。我们将此类系统命名为"数据感知NAS"，并通过与传统NAS的对比实验，为其优越性提供了实证依据。