Analog Nanoscale Electro-Optical Synapses for Neuromorphic Computing Applications
Source abstract
Abstract
The typically nonlinear and asymmetric response of synaptic memristors to positive and negative electrical pulses makes the realization of accurate deep neural networks very challenging. Here, we integrate a two-terminal valence change memory (VCM) into a photonic/plasmonic circuit and show that the switching properties of this memristor become more gradual and symmetric under light irradiation. The added optical input acts on the VCM as a third, independent modulation channel. It locally heats the active area of the device, which enhances the generation of oxygen vacancies and broadens the resulting nanoscale conductive filaments. The measured conductance modulation of the VCM is then inserted into a neural network simulator. Using the MNIST data set of handwritten digits as an application, a light-enhanced recognition accuracy of 93.53% is demonstrated, similar to ideally performing memristors (94.86%) and much higher than those without light (67.37%). Notably, the optical signal does not increase the overall energy consumption by more than 3.2%. Finally, an approach to scale up our electro-optical technology is proposed, which could allow high-density, energy-efficient neuromorphic computing chips.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Advanced Memory and Neural Computing · Neural Networks and Reservoir Computing · Photoreceptor and optogenetics research
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Integrated Systems Laboratory, ETH Zurich, 8092 Zurich, Switzerland
Institute of Electromagnetic Fields (IEF), ETH Zurich, 8092 Zurich, Switzerland
imec, B-3001 Leuven, Belgium
Funding
ETH Z??rich Foundation · Werner Siemens Stiftung · Swiss National Supercomputing Centre
References (47)
Show source-provided references
1. 10.1038/s42254-020-0208-210.1038/s42254-020-0208-2 ↗
2. 10.1109/5.5835610.1109/5.58356 ↗
3. 10.1080/23746149.2016.125958510.1080/23746149.2016.1259585 ↗
4. 10.1038/nature1453910.1038/nature14539 ↗
5. Coates A. · Proc. Mach. Learn. Res. · 2013 · 1337
6. 10.1145/3079856.308024610.1145/3079856.3080246 ↗
7. 10.1109/JPROC.2017.276174010.1109/JPROC.2017.2761740 ↗
8. 10.1109/TCT.1971.108333710.1109/TCT.1971.1083337 ↗
9. 10.1038/nature0693210.1038/nature06932 ↗
10. 10.1038/s41586-020-1942-410.1038/s41586-020-1942-4 ↗
11. 10.1038/nature1444110.1038/nature14441 ↗
12. 10.1038/s41586-018-0180-510.1038/s41586-018-0180-5 ↗
13. 10.1109/JPROC.2018.279084010.1109/JPROC.2018.2790840 ↗
14. 10.1109/TED.2015.243963510.1109/TED.2015.2439635 ↗
15. 10.1109/ICCAD.2015.737257010.1109/ICCAD.2015.7372570 ↗
16. 10.1109/IJCNN.2016.772729810.1109/IJCNN.2016.7727298 ↗
17. 10.3389/fnins.2016.0033310.3389/fnins.2016.00333 ↗
18. 10.1063/1.493481810.1063/1.4934818 ↗
19. 10.23919/VLSIT.2019.877655110.23919/VLSIT.2019.8776551 ↗
20. 10.1109/IEDM.2017.826837210.1109/IEDM.2017.8268372 ↗
21. 10.1002/adma.20160431010.1002/adma.201604310 ↗
22. 10.1038/nmat485610.1038/nmat4856 ↗
23. 10.1109/LED.2016.258285910.1109/LED.2016.2582859 ↗
24. 10.1002/aelm.20170003210.1002/aelm.201700032 ↗
25. 10.1038/s41928-017-0002-z10.1038/s41928-017-0002-z ↗
26. 10.1038/ncomms367610.1038/ncomms3676 ↗
27. 10.1038/s41427-019-0182-210.1038/s41427-019-0182-2 ↗
28. 10.1038/nphoton.2015.18210.1038/nphoton.2015.182 ↗
29. 10.1126/sciadv.170016010.1126/sciadv.1700160 ↗
30. 10.1038/s41586-020-03070-110.1038/s41586-020-03070-1 ↗
31. 10.1021/acsnano.8b0181110.1021/acsnano.8b01811 ↗
32. 10.1021/acs.nanolett.5b0453710.1021/acs.nanolett.5b04537 ↗
33. 10.1109/IEDM.2017.826832610.1109/IEDM.2017.8268326 ↗
34. 10.1038/323533a010.1038/323533a0 ↗
35. 10.1016/j.sse.2016.07.00610.1016/j.sse.2016.07.006 ↗
36. 10.1109/5.72679110.1109/5.726791 ↗
37. 10.1038/nphoton.2009.2610.1038/nphoton.2009.26 ↗
38. 10.1109/IEDM.2017.826837410.1109/IEDM.2017.8268374 ↗
39. 10.1109/TED.2014.232553110.1109/TED.2014.2325531 ↗
40. 10.1364/OE.18.01859210.1364/OE.18.018592 ↗
41. 10.1109/IEDM.2016.783846510.1109/IEDM.2016.7838465 ↗
42. 10.1109/LED.2010.204189310.1109/LED.2010.2041893 ↗
43. 10.1109/IEDM.2017.826833710.1109/IEDM.2017.8268337 ↗
44. Abadi M. · OSDI ’16: Proc. 12th USENIX Symp. Oper. Syst. Des. Implementation · 2016 · 265
45. 10.1109/IJCNN.2015.728078510.1109/IJCNN.2015.7280785 ↗
46. 10.3389/fncir.2015.0008510.3389/fncir.2015.00085 ↗
47. 10.1063/5.002853910.1063/5.0028539 ↗