Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat
Cited by 253

The DOI matches its Crossref record; 218 of the 218 references it cites were found in a registry or an index.

Machine Learning for Microcontroller-Class Hardware: A Review

Open access
Saha, Swapnil Sayan, Sandha, Sandeep Singh, Srivastava, Mani · 2022 · IEEE Sensors Journal

Abstract

The advancements in machine learning opened a new opportunity to bring intelligence to the low-end Internet-of-Things nodes such as microcontrollers. Conventional machine learning deployment has high memory and compute footprint hindering their direct deployment on ultra resource-constrained microcontrollers. This paper highlights the unique requirements of enabling onboard machine learning for microcontroller class devices. Researchers use a specialized model development workflow for resource-limited applications to ensure the compute and latency budget is within the device limits while still maintaining the desired performance. We characterize a closed-loop widely applicable workflow of machine learning model development for microcontroller class devices and show that several classes of applications adopt a specific instance of it. We present both qualitative and numerical insights into different stages of model development by showcasing several use cases. Finally, we identify the open research challenges and unsolved questions demanding careful considerations moving forward.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1109/jsen.2022.3210773

Cites

218 of 218 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Items 201–218 of 218

Research works printed but not linked: 52

  1. 2022, X-CUBE-AI—AI Expansion Pack for STM32CubeMX—Stmicroelectronics
  2. goyal, 2020, DROCC: Deep robust one-class classification, Proc Int Conf Mach Learn, 3711
  3. morawiec, 2022, sklearn-porter
  4. 2021, Nanoedge AI Library
  5. 2022, Edge Ai/TinyML SAAS Deep Learning
  6. 2022, AI-Driven Deep Neural Network Optimizer
  7. tan, 2022, Utensor/Utensor TinyML Ai Inference Library
  8. dennis, 2019, Shallow RNN: Accurate time-series classification on resource constrained devices, Proc Adv Neural Inf Process Syst, 32, 1
  9. 2022, Edge Impulse
  10. 2022, Plumerai
  11. 2022, Qeexo Automl
  12. 2022, Sensiml Making Sensor Data Sensible
  13. 2022, Reality AI
  14. 2022, Neuton AI—No-Code Artificial Intelligence for All
  15. wang, 2019, Picking winning tickets before training by preserving gradient flow, Proc Int Conf Learn Represent, 1
  16. chen, 2020, Neural architecture search on ImageNet in four GPU hours: A theoretically inspired perspective, Proc Int Conf Learn Represent, 1
  17. tanaka, 2020, Pruning neural networks without any data by iteratively conserving synaptic flow, Proc Adv Neural Inf Process Syst, 33, 6377
  18. mellor, 2021, Neural architecture search without training, Proc Int Conf Mach Learn, 7588
  19. 10.1109/ICASSP40776.2020.9054609 · DOI
  20. daulton, 2022, Multi-objective Bayesian optimization over high-dimensional search spaces, Proc Conf Uncertainty of Artificial Intelligence, 507
  21. gupta, 2017, ProtoNN: Compressed and accurate KNN for resource-scarce devices, Proc Int Conf Mach Learn, 1331
  22. 2022, Microsoft/Nni An Open Source Automl Toolkit for Automate Machine Learning Lifecycle Including Feature Engineering Neural Architecture Search Model Compression and Hyper-Parameter Tuning
  23. bannink, 2021, Larq compute engine: Design, benchmark and deploy state-of-the-art binarized neural networks, Proc Mach Learn Syst, 3, 680
  24. tauzin, 2021, giotto-tda: A topological data analysis toolkit for machine learning and data exploration, J Mach Learn Res, 22, 1
  25. 10.1007/978-1-4615-4337-4 · doi:10.1007/978-1-4615-4337-4
  26. gielda, 2017, Renode: A flexible, open-source simulation framework for building scalable, well-tested RISC-V systems, Proc 7th RISC-V Workshop, 1
  27. jung, 2019, Imgaug documentation, Readthedocs io, 25, 1
  28. clark, 2015, Pillow (pil fork) documentation, Readthedocs io, 6 2 0, 1
  29. 2022, Eloquentarduino
  30. han, 2015, Learning both weights and connections for efficient neural network, Proc Adv Neural Inf Process Syst, 28, 1
  31. mcmahan, 2017, Communication-efficient learning of deep networks from decentralized data, Proc 20th Int Conf Artif Intell Statist, 1273
  32. 10.1109/VLSID.2019.00056 · DOI
  33. 10.1109/CVPR42600.2020.01079 · doi:10.1109/cvpr42600.2020.01079 · Paper page
  34. hubara, 2016, Binarized neural networks, Proc Adv Neural Inf Process Syst, 29, 1
  35. rusci, 2020, Memory-driven mixed low precision quantization for enabling deep network inference on microcontrollers, Proc Mach Learn Syst, 2, 326
  36. liberis, 2022, Differentiable network pruning to enable smart applications, Proc 4th U K Mobile Wearable Ubiquitous Syst Res Symp, 1
  37. 10.1109/ET52713.2021.9579991 · DOI
  38. baker, 2017, Designing neural network architectures using reinforcement learning, Proc Int Conf Learn Represent (ICLR), 1
  39. cai, 2018, ProxylessNAS: Direct neural architecture search on target task and hardware, Proc Int Conf Learn Represent, 1
  40. chen, 2018, Understanding the limitations of existing energy-efficient design approaches for deep neural networks, Energy, 2, 3l
  41. cranmer, 2020, Lagrangian neural networks, Proc ICLR WKSH Integr Deep Neural Models Differ Equ, 1
  42. mao, 2018, The neuro-symbolic concept learner: Interpreting scenes, words, and sentences from natural supervision, Proc Int Conf Learn Represent, 1
  43. li, 2021, FLASH: Fast neural architecture search with hardware optimization, ACM Trans Embedded Comput Syst, 20, 1, 10.1145/3477034 · doi:10.1145/3477034
  44. silver, 2014, Deterministic policy gradient algorithms, Proc Int Conf Mach Learn, 387
  45. elsken, 2018, Efficient multi-objective neural architecture search via Lamarckian evolution, Proc Int Conf Learn Represent, 1
  46. jose, 2018, Kronecker recurrent units, Proc Int Conf Mach Learn, 2380
  47. vaswani, 0, Attention is all you need, Proc Adv Neural Inf Process Syst, 2017, 5998
  48. lin, 2018, Defensive quantization: When efficiency meets robustness, Proc Int Conf Learn Represent, 1
  49. jagielski, 2020, High accuracy and high fidelity extraction of neural networks, Proc 29th USENIX Secur Symp (USENIX Secur ), 1345
  50. gui, 2019, Model compression with adversarial robustness: A unified optimization framework, Proc Adv Neural Inf Process Syst, 32, 1285
  51. cai, 2020, TinyTL: Reduce memory, not parameters for efficient on-device learning, Proc Adv Neural Inf Process Syst, 33, 1
  52. 10.3390/s21144805 · DOI

Cited by

See all 253

Metadata

Authors
Saha, S. S., Sandha, S. S., & Srivastava, M.
Venue
IEEE Sensors Journal
Published
2022-11-15
Volume
22
Issue
22
Pages
21362-21390
ISSN
  • 1530-437X (ISSN-L)
  • 1558-1748
  • 2379-9153
OpenAlex
W4309287541
Cited by
253
References
218
FWCI
21.17
Type
journal article
Language
en
Open access
green
Funders
  • National Institutes of Health (1P41EB028242)
  • Army Research Laboratory (W911NF-17-2-0196)
  • Semiconductor Research Corporation (S201617)

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.