# Neural Architecture Search for Energy Efficient Always-on Audio Models

arXiv: https://arxiv.org/abs/2202.05397 · HTML: https://mlsanity.com/paper/2202.05397/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 29 references that are in the map.
- First posted on arXiv 9 Feb 2022.

## Builds on (most cited first, 18 of 18)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://mlsanity.com/paper/1801.04381/) (2018)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [Searching for MobileNetV3](https://mlsanity.com/paper/1905.02244/) (2019)
- [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://mlsanity.com/paper/1904.08779/) (2019)
- [ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices](https://mlsanity.com/paper/1707.01083/) (2017)
- [MnasNet: Platform-Aware Neural Architecture Search for Mobile](https://mlsanity.com/paper/1807.11626/) (2018)
- [Regularized Evolution for Image Classifier Architecture Search](https://mlsanity.com/paper/1802.01548/) (2018)
- [CNN Architectures for Large-Scale Audio Classification](https://mlsanity.com/paper/1609.09430/) (2016)
- [ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware](https://mlsanity.com/paper/1812.00332/) (2018)
- [FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search](https://mlsanity.com/paper/1812.03443/) (2018)
- [Random Search and Reproducibility for Neural Architecture Search](https://mlsanity.com/paper/1902.07638/) (2019)
- [NetAdapt: Platform-Aware Neural Network Adaptation for Mobile Applications](https://mlsanity.com/paper/1804.03230/) (2018)
- [Freeze-Thaw Bayesian Optimization](https://mlsanity.com/paper/1406.3896/) (2014)
- [Can weight sharing outperform random architecture search? An investigation with TuNAS](https://mlsanity.com/paper/2008.06120/) (2020)
- [Temporal Convolution for Real-time Keyword Spotting on Mobile Devices](https://mlsanity.com/paper/1904.03814/) (2019)
- [One Proxy Device Is Enough for Hardware-Aware Neural Architecture Search](https://mlsanity.com/paper/2111.01203/) (2021)
- [Resource-Aware Pareto-Optimal Automated Machine Learning Platform](https://mlsanity.com/paper/2011.00073/) (2020)

## Built on by (most cited first, 5 of 5)

- [Green Edge AI: A Contemporary Survey](https://mlsanity.com/paper/2312.00333/) (2023)
- [Band-gap regression with architecture-optimized message-passing neural networks](https://mlsanity.com/paper/2309.06348/) (2023)
- [Training speedups via batching for geometric learning: an analysis of static and dynamic algorithms](https://mlsanity.com/paper/2502.00944/) (2025)
- [Extrapolation to complete basis-set limit in density-functional theory by quantile random-forest models](https://mlsanity.com/paper/2303.14760/) (2023)
- [Scaling Up Quantization-Aware Neural Architecture Search for Efficient Deep Learning on the Edge](https://mlsanity.com/paper/2401.12350/) (2024)

## Abstract (arXiv)

Mobile and edge computing devices for always-on classification tasks require energy-efficient neural network architectures. In this paper we present several changes to neural architecture searches (NAS) that improve the chance of success in practical situations. Our search simultaneously optimizes for network accuracy, energy efficiency and memory usage. We benchmark the performance of our search on real hardware, but since running thousands of tests with real hardware is difficult we use a random forest model to roughly predict the energy usage of a candidate network. We present a search strategy that uses both Bayesian and regularized evolutionary search with particle swarms, and employs early-stopping to reduce the computational burden. Our search, evaluated on a sound-event classification dataset based upon AudioSet, results in an order of magnitude less energy per inference and a much smaller memory footprint than our baseline MobileNetV1/V2 implementations while slightly improving task accuracy. We also demonstrate how combining a 2D spectrogram with a convolution with many filters causes a computational bottleneck for audio classification and that alternative approaches reduce the computational burden but sacrifice task accuracy.