mlsanityData to 7 Oct 2026

Sound Event Detection with Binary Neural Networks on Tightly Power-Constrained IoT Devices

arXiv 2101.04446 ↗First posted 12 Jan 2021cs.LG eess.AScited by 10

Not on the map. Why?

Abstract (arXiv)

Sound event detection (SED) is a hot topic in consumer and smart city applications. Existing approaches based on Deep Neural Networks are very effective, but highly demanding in terms of memory, power, and throughput when targeting ultra-low power always-on devices. Latency, availability, cost, and privacy requirements are pushing recent IoT systems to process the data on the node, close to the sensor, with a very limited energy supply, and tight constraints on the memory size and processing capabilities precluding to run state-of-the-art DNNs. In this paper, we explore the combination of extreme quantization to a small-footprint binary neural network (BNN) with the highly energy-efficient, RISC-V-based (8+1)-core GAP8 microcontroller. Starting from an existing CNN for SED whose footprint (815 kB) exceeds the 512 kB of memory available on our platform, we retrain the network using binary filters and activations to match these memory constraints. (Fully) binary neural networks come with a natural drop in accuracy of 12-18% on the challenging ImageNet object recognition challenge compared to their equivalent full-precision baselines. This BNN reaches a 77.9% accuracy, just 7% lower than the full-precision version, with 58 kB (7.2 times less) for the weights and 262 kB (2.4 times less) memory in total. With our BNN implementation, we reach a peak throughput of 4.6 GMAC/s and 1.5 GMAC/s over the full network, including preprocessing with Mel bins, which corresponds to an efficiency of 67.1 GMAC/s/W and 31.3 GMAC/s/W, respectively. Compared to the performance of an ARM Cortex-M4 implementation, our system has a 10.3 times faster execution time and a 51.1 times higher energy-efficiency.

Key facts. Cited by 10 papers in the map (data through 7 Oct 2026). Builds on 19 of its 42 references that are in the map. First posted on arXiv 12 Jan 2021.

Builds on

19 of 42 references.

References · 19

  1. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  2. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,548
  3. 2016SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model sizecited by 1,054
  4. 2016XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networkscited by 829
  5. 2016CNN Architectures for Large-Scale Audio Classificationcited by 701
  6. 2017Channel Pruning for Accelerating Very Deep Neural Networkscited by 697
  7. 2016DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradientscited by 557
  8. 2015Fixed Point Quantization of Deep Convolutional Networkscited by 141
  9. 2018CMSIS-NN: Efficient Neural Network Kernels for Arm Cortex-M CPUscited by 116
  10. 2017Hello Edge: Keyword Spotting on Microcontrollerscited by 108
  11. 2019FastGRNN: A Fast, Accurate, Stable and Tiny Kilobyte Sized Gated Recurrent Neural Networkcited by 38
  12. 2018Structured Binary Neural Networks for Accurate Image Classification and Semantic Segmentationcited by 37
7 more
  1. 2019PULP-NN: Accelerating Quantized Neural Networks on Parallel Ultra-Low-Power RISC-V Processorscited by 34
  2. 2018A 64mW DNN-based Visual Navigation Engine for Autonomous Nano-Dronescited by 21
  3. 2019FANN-on-MCU: An Open-Source Toolkit for Energy-Efficient Neural Network Inference at the Edge of the Internet of Thingscited by 17
  4. 2018BinarEye: An Always-On Energy-Accuracy-Scalable Binary CNN Processor With All Memory On Chip in 28nm CMOScited by 10
  5. 2020Compact recurrent neural networks for acoustic event detection on low-energy low-complexity platformscited by 6
  6. 2017Efficient Convolutional Neural Network For Audio Event Detectioncited by 5
  7. 2019Additive Noise Annealing and Approximation Properties of Quantized Neural Networkscited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 2; 2022: 1; 2023: 3; 2024: 0; 2025: 2; 2026: 2.

10 papers.

  1. 2023TinyTracker: Ultra-Fast and Ultra-Low-Power Edge Vision In-Sensor for Gaze Estimationcited by 16
  2. 2025Neural Network Quantization for Microcontrollers: A Comprehensive Survey of Methods, Platforms, and Applicationscited by 6
  3. 2021Environmental Sound Classification on the Edge: A Pipeline for Deep Acoustic Networks on Extremely Resource-Constrained Devicescited by 5
  4. 2022Sub-mW Keyword Spotting on an MCU: Analog Binary Feature Extraction and Binary Neural Networkscited by 5
  5. 2021Pruning vs XNOR-Net: A Comprehensive Study of Deep Learning for Audio Classification on Edge-devicescited by 3
  6. 2023Enhancing Lightweight Neural Networks for Small Object Detection in IoT Applicationscited by 2
  7. 2023FMAS: Fast Multi-Objective SuperNet Architecture Search for Semantic Segmentationcited by 0
  8. 2025MUDAS: Mote-scale Unsupervised Domain Adaptation in Multi-label Sound Classificationcited by 0
  9. 2026Active Continual Learning with Metaplastic Binary Bayesian Neural Networkscited by 0
  10. 2026A Neuromorphic Trigger for Efficient Audio Event Detectioncited by 0