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

The DOI matches its Crossref record; 41 of the 45 references it cites were found in a registry or an index; 6 of the 26 sources its claims cite were checked for support.

Powering AI at the edge: A robust, memristor-based binarized neural network with near-memory computing and miniaturized solar cell

Open access

Abstract

Memristor-based neural networks provide an exceptional energy-efficient platform for artificial intelligence (AI), presenting the possibility of self-powered operation when paired with energy harvesters. However, most memristor-based networks rely on analog in-memory computing, necessitating a stable and precise power supply, which is incompatible with the inherently unstable and unreliable energy harvesters. In this work, we fabricated a robust binarized neural network comprising 32,768 memristors, powered by a miniature wide-bandgap solar cell optimized for edge applications. Our circuit employs a resilient digital near-memory computing approach, featuring complementarily programmed memristors and logic-in-sense-amplifier. This design eliminates the need for compensation or calibration, operating effectively under diverse conditions. Under high illumination, the circuit achieves inference performance comparable to that of a lab bench power supply. In low illumination scenarios, it remains functional with slightly reduced accuracy, seamlessly transitioning to an approximate computing mode. Through image classification neural network simulations, we demonstrate that misclassified images under low illumination are primarily difficult-to-classify cases. Our approach lays the groundwork for self-powered AI and the creation of intelligent sensors for various applications in health, safety, and environment monitoring.

Overview

CoChat Summary

## TL;DR

The paper presents a hybrid CMOS/memristor binarized neural network that is designed to keep working under unstable, harvested power rather than requiring a tightly regulated supply. In experiments, it functioned down to 0.7 V and under illumination as low as 0.08 suns, with accuracy degrading gracefully rather than failing outright.

## Key findings

  • The fabricated system integrates 32,768 memristors and operates as a robust binarized neural network using a differential, logic-in-sense-amplifier design that the authors say removes the need for compensation or calibration.
  • With a lab-bench supply, the system had no errors at ≥1.0 V; at 0.9 V, occasional errors appeared at 66 MHz, and below that errors reached up to 2% at any frequency. The lowest measured energy consumption was 45 nJ at 0.7 V and 10 MHz.
  • Under harvested power, the circuit remained functional even at an equivalent solar power of 0.08 suns. For MNIST, accuracy dropped by only 0.7 percentage points at 0.08 suns, while for CIFAR-10 accuracy fell from 86.6% to 73.4%.
  • The authors report an energy efficiency of 2.9 TOPS/W under optimal fabricated conditions after removing digital control energy; they estimate 22.5 TOPS/W after further optimizing away clock distribution and register energy, and project 397 TOPS/W for a scaled 28-nm design.

## Methods

  • Design: A hybrid CMOS/nanotechnology binarized neural network with distributed memristor memory modules, complementary programming of two memristors per weight, and near-memory population-count accumulation.
  • Data/sample: Hardware experiments used a fabricated demonstrator and a miniature AlGaAs/GaInP wide-bandgap solar cell; neural-network simulations were run for MNIST and CIFAR-10.
  • Main analysis: Electrical characterization was done across supply voltages and frequencies with lab equipment; solar-powered inference was tested by directly connecting the chip to the solar cell; network accuracy under hardware errors was evaluated by injecting experimentally measured error rates into PyTorch simulations.

## Limitations

  • The paper states that the current demonstrator handles single-layer computations in hardware; multilayer processing, input buffering, and convolution are not embedded on-chip and would be needed in a final commercial chip.
  • The solar cell used had 8.7% efficiency under AM1.5G, and the authors explicitly note this was limited by an unoptimized process and no anti-reflection coating; they say similar stacks reached 18.7% with optimization.
  • Evident from the text: the fabricated chip was measured on a packaged version with external lab equipment, so the system is not yet a fully integrated standalone self-powered device.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41467-024-44766-6

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Overview of the fabricated memristor-based binarized neural network. a Optical microscopy image of the fabricated die, showing four memory modules and their associated digital circuitry and power management unit. b Detail on one of the memory modules. c Cross-sectional scanning electron micrograph of a hybrid CMOS/memristor circuit, showing a memristor between metal levels four and five. d Schematic of a memory module. For each operation mode, biasing conditions for WL, BL, and SL are given with respect to the power domain (VDDC, VDDR) and VDD. e Schematic of the level shifter, used in d for shifting digital voltage input to medium voltages needed during programming operations or nominal voltage during reading operations of the memristors. f Schematic of the differential pre-charge sense amplifier PCSA, used to read the binary memristor states, with embedded XNOR function, to compose a XPCSA: it computes an XNOR operation between input activation X and weight (memristor value) during bit-cell sensing.
Fig. 1 Overview of the fabricated memristor-based binarized neural network. a Optical microscopy image of the fabricated die, showing four memory modules and their associated digital circuitry and power management unit. b Detail on one of the memory modules. c Cross-sectional scanning electron micrograph of a hybrid CMOS/memristor circuit, showing a memristor between metal levels four and five. d Schematic of a memory module. For each operation mode, biasing conditions for WL, BL, and SL are given with respect to the power domain (VDDC, VDDR) and VDD. e Schematic of the level shifter, used in d for shifting digital voltage input to medium voltages needed during programming operations or nominal voltage during reading operations of the memristors. f Schematic of the differential pre-charge sense amplifier PCSA, used to read the binary memristor states, with embedded XNOR function, to compose a XPCSA: it computes an XNOR operation between input activation X and weight (memristor value) during bit-cell sensing.CC BY · PubMed Central
Fig. 2 Measurements of the memristor-based binarized neural network, employing a lab-bench power supply. a Sample measurement of the output of the integrated circuit, compared with a delay-less register-transfer level (RTL) simulation. b Photograph of the printed circuit board used for the experiments. c Measurement of the energy consumption to perform a whole-chip inference, for various operating frequencies and supply voltages. d Pie chart comparing the different sources of energy consumption in the system, obtained using simulations (see Methods).
Fig. 2 Measurements of the memristor-based binarized neural network, employing a lab-bench power supply. a Sample measurement of the output of the integrated circuit, compared with a delay-less register-transfer level (RTL) simulation. b Photograph of the printed circuit board used for the experiments. c Measurement of the energy consumption to perform a whole-chip inference, for various operating frequencies and supply voltages. d Pie chart comparing the different sources of energy consumption in the system, obtained using simulations (see Methods).CC BY · PubMed Central
Fig. 3 Accuracy of the memristor-based binarized neural network. a Measured schmoo plot, presenting mean accuracy of the output neuron activations, for different operation frequency and supply voltage. They were obtained using patterns of weights and inputs chosen to cover all possible neuron preactivations (see Methods). NF means non-functional. b Measurements of 64 neurons with preactivations –5, –1, 0, 1, and 5, at 33 and 66 MHz with a power supply of 0.9 volts. Errors are marked in red. Mean accuracy of the output neuron activations, as a function of neuron preactivation Δ and supply voltage, measured at (c) 33 and (d) 66 MHz (see Methods).
Fig. 3 Accuracy of the memristor-based binarized neural network. a Measured schmoo plot, presenting mean accuracy of the output neuron activations, for different operation frequency and supply voltage. They were obtained using patterns of weights and inputs chosen to cover all possible neuron preactivations (see Methods). NF means non-functional. b Measurements of 64 neurons with preactivations –5, –1, 0, 1, and 5, at 33 and 66 MHz with a power supply of 0.9 volts. Errors are marked in red. Mean accuracy of the output neuron activations, as a function of neuron preactivation Δ and supply voltage, measured at (c) 33 and (d) 66 MHz (see Methods).CC BY · PubMed Central
Fig. 4 Measurements of the binarized neural network powered by a miniature solar cell. a Schematic view of the AlGaAs/GaInP heterostructure solar cell. b Photograph of the solar cell, and its measured current-voltage characteristics under one-sun AM1.5 illumination provided by a standardized solar simulator (see Methods). c Current-voltage characteristics of the solar cell for various illuminations provided by the halogen lamp (see Methods). d Photograph of the experimental setup where the fabricated binarized neural network is powered by the solar cell illuminated by the halogen lamp. e Mean measured accuracy of the output neuron activations, with the binarized neural network powered by the solar cell, as a function of neuron preactivation Δ and solar cell illumination.
Fig. 4 Measurements of the binarized neural network powered by a miniature solar cell. a Schematic view of the AlGaAs/GaInP heterostructure solar cell. b Photograph of the solar cell, and its measured current-voltage characteristics under one-sun AM1.5 illumination provided by a standardized solar simulator (see Methods). c Current-voltage characteristics of the solar cell for various illuminations provided by the halogen lamp (see Methods). d Photograph of the experimental setup where the fabricated binarized neural network is powered by the solar cell illuminated by the halogen lamp. e Mean measured accuracy of the output neuron activations, with the binarized neural network powered by the solar cell, as a function of neuron preactivation Δ and solar cell illumination.CC BY · PubMed Central
Fig. 5 Neural-network-level investigations. a, b Illustration of our method for mapping arbitrarily-shaped binarized neural networks to 64 × 128 memristor arrays. The detailed methodology is presented in Supplementary Note 4. c t-distributed stochastic neighbor embedding (t-SNE) representation of the MNIST test dataset. The datapoints incorrectly classified under 0.8 suns (left) and 0.08 suns (right) equivalent illumination, but which would be correctly classified under 8 suns, are marked in black. These results are obtained using a binarized fully-connected neural network (see Methods). d Same graphs as c, obtained for the CIFAR-10 dataset and using a convolutional neural network (see Methods).
Fig. 5 Neural-network-level investigations. a, b Illustration of our method for mapping arbitrarily-shaped binarized neural networks to 64 × 128 memristor arrays. The detailed methodology is presented in Supplementary Note 4. c t-distributed stochastic neighbor embedding (t-SNE) representation of the MNIST test dataset. The datapoints incorrectly classified under 0.8 suns (left) and 0.08 suns (right) equivalent illumination, but which would be correctly classified under 8 suns, are marked in black. These results are obtained using a binarized fully-connected neural network (see Methods). d Same graphs as c, obtained for the CIFAR-10 dataset and using a convolutional neural network (see Methods).CC BY · PubMed Central

Verification

Claims the paper makes

5 Valid · 1 Not valid · 12 Not checked · 90 Own results and interpretations

55 more claims

Create a free account to see all 108 claims and to verify the 5 references still open.

See all claims — free account
  • It may state two separate claims as one. 37
  • Probably not a substantial claim. 15

Create a free account to see these claims.

Citations

Cites

Show all 45 references

Cited by

See all 58

Metadata

Authors
Jebali, F., Majumdar, A., Turck, C., Harabi, K.-E., Faye, M.-C., Muhr, E., Walder, J.-P., Bilousov, O., Michaud, A., Vianello, E., Hirtzlin, T., Andrieu, F., Bocquet, M., Collin, S., Querlioz, D., & Portal, J.-M.
Affiliations
  • Centre National de la Recherche Scientifique (FR)
  • Centre de Nanosciences et de Nanotechnologies (FR)
  • Commissariat à l'Énergie Atomique et aux Énergies Alternatives (FR)
Corresponding author
  • Querlioz, Damien
  • Portal, Jean-Michel
Venue
Nature Communications
Published
2024-01-25
Volume
15
Issue
1
Pages
741-741
ISSN
2041-1723 (ISSN-L)
OpenAlex
W4391222208
Cited by
58
References
45
Citations by year
2026: 23 · 2025: 26 · 2024: 8
FWCI
5.88
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • HAL AMU (submittedVersion, cc-by)
  • HAL AMU (submittedVersion, other-oa)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • arXiv (Cornell University)
Funders
  • European Commission (H2020, 101007321)
  • Agence Nationale de la Recherche (France2030, ANRIEED-002-01, ANR18, ANR-18-CE24-0009, ANR-22-PEEL-0010, IEED-002-01)
  • Electronic Components and Systems for European Leadership (101007321)
  • Horizon 2020 Framework Programme (101007321)
  • H2020 European Research Council (715872)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Memory and Neural Computing
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Low-power high-performance VLSI design
  • Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Reservoir Computing
Keywords
memristor-based neural networks, binary neural networks, near-memory computing, energy harvesting, edge computing, approximate computing, image classification, smart sensors

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Powering AI at the edge: A robust, memristor-based binarized neural network with near-memory computing and miniaturized solar cell. https://app.cochat.ai/papers/10.1038/s41467-024-44766-6
The paper
Jebali, F., Majumdar, A., Turck, C., Harabi, K.-E., Faye, M.-C., Muhr, E., Walder, J.-P., Bilousov, O., Michaud, A., Vianello, E., Hirtzlin, T., Andrieu, F., Bocquet, M., Collin, S., Querlioz, D., & Portal, J.-M. (2024). Powering AI at the edge: A robust, memristor-based binarized neural network with near-memory computing and miniaturized solar cell. Nature Communications. https://doi.org/10.1038/s41467-024-44766-6

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.