A Cascaded Duplex Organic Vertical Memory with Learning Rate Scheduling for Efficient Artificial Neural Network Training
Source abstract
Abstract
Abstract Learning rate scheduling (LRS) is a critical factor influencing the performance of neural networks by accelerating the convergence of learning algorithms and enhancing the generalization capabilities. The escalating computational demands in artificial intelligence (AI) necessitate advanced hardware solutions capable of supporting neural network training with LRS. This not only requires linear and symmetric analog programming capabilities but also the precise adjustment of channel conductance to achieve tunable slope in weight update behaviors. Here, a cascaded duplex organic vertical memory is proposed with the coupling of ferroelectric polarization effect and Schottky gate control on the same semiconducting channel, exhibiting adjustable‐slope conductance update with high linearity and symmetry. Therefore, in the chest X‐ray image detection, a fast‐to‐slow LRS is used for a bi‐layer ANN training, achieving a rapid, stable convergence behavior within only 15 epochs and a high recognition accuracy. Moreover, the proposed LRS training is also suitable for the Mackey Glass prediction task using long short‐term memory networks. This work integrates LRS into synaptic devices, enabling efficient hardware implementation of neural networks and thus enhancing AI performance in practical applications.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Advanced Memory and Neural Computing · Ferroelectric and Negative Capacitance Devices · Neuroscience and Neural Engineering
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
National Laboratory of Solid‐State Microstructures School of Electronic Science and Engineering Collaborative Innovation Center of Advanced Microstructures Nanjing University Nanjing 210093 P. R. China
Funding
National Key Research and Development Program of China · National Natural Science Foundation of China
References (47)
Show source-provided references
1. 10.1038/s41563-019-0291-x10.1038/s41563-019-0291-x ↗
2. 10.1038/s41565-023-01343-010.1038/s41565-023-01343-0 ↗
3. 10.1038/s41565-020-0655-z10.1038/s41565-020-0655-z ↗
4. 10.1038/s41586-021-04362-w10.1038/s41586-021-04362-w ↗
5. 10.1038/s41586-020-1942-410.1038/s41586-020-1942-4 ↗
6. 10.1109/JSAC.2021.311840210.1109/JSAC.2021.3118402 ↗
7. 10.1109/JIOT.2022.321984710.1109/JIOT.2022.3219847 ↗
8. 10.1109/ACCESS.2024.341304310.1109/ACCESS.2024.3413043 ↗
9. 10.1038/s41467-022-29411-410.1038/s41467-022-29411-4 ↗
10. 10.1038/s41928-019-0221-610.1038/s41928-019-0221-6 ↗
11. 10.1038/s41928-022-00878-910.1038/s41928-022-00878-9 ↗
12. 10.3390/electronics1223480310.3390/electronics12234803 ↗
13. 10.1038/s41467-023-36270-010.1038/s41467-023-36270-0 ↗
14. 10.1002/adma.20221062810.1002/adma.202210628 ↗
15. 10.1038/s41467-018-04484-210.1038/s41467-018-04484-2 ↗
16. 10.1038/s42256-019-0089-110.1038/s42256-019-0089-1 ↗
17. 10.1038/s41928-022-00876-x10.1038/s41928-022-00876-x ↗
18. 10.1002/adma.20200398410.1002/adma.202003984 ↗
19. 10.1038/s41467-023-41958-410.1038/s41467-023-41958-4 ↗
20. 10.1126/science.abp806410.1126/science.abp8064 ↗
21. 10.1038/s41928-019-0316-010.1038/s41928-019-0316-0 ↗
22. 10.1002/aelm.20230039510.1002/aelm.202300395 ↗
23. 10.1002/adma.20190296210.1002/adma.201902962 ↗
24. 10.1002/adfm.20201030310.1002/adfm.202010303 ↗
25. 10.1126/sciadv.abf874410.1126/sciadv.abf8744 ↗
26. 10.1038/s41928-020-0435-710.1038/s41928-020-0435-7 ↗
27. 10.1038/s41467-021-27575-z10.1038/s41467-021-27575-z ↗
28. 10.1038/s41467-019-11411-610.1038/s41467-019-11411-6 ↗
29. 10.1002/adma.20220686410.1002/adma.202206864 ↗
30. 10.1039/D3MH00429E10.1039/D3MH00429E ↗
31. 10.1038/s41467-018-07565-410.1038/s41467-018-07565-4 ↗
32. 10.1038/s41467-022-31804-410.1038/s41467-022-31804-4 ↗
33. 10.1080/00018732.2022.208400610.1080/00018732.2022.2084006 ↗
34. 10.1039/D0MH02029J10.1039/D0MH02029J ↗
35. 10.1038/nature1067910.1038/nature10679 ↗
36. 10.1109/TCAD.2018.278972310.1109/TCAD.2018.2789723 ↗
37. 10.1109/TCAD.2020.304373110.1109/TCAD.2020.3043731 ↗
38. 10.1021/acs.nanolett.9b0018010.1021/acs.nanolett.9b00180 ↗
39. 10.1038/s41565-020-0694-510.1038/s41565-020-0694-5 ↗
40. 10.1109/JPROC.2018.287105710.1109/JPROC.2018.2871057 ↗
41. 10.3389/fnins.2016.0033310.3389/fnins.2016.00333 ↗
42. 10.3389/fnins.2020.0010310.3389/fnins.2020.00103 ↗
43. 10.1038/s41467-018-07904-510.1038/s41467-018-07904-5 ↗
44. 10.1016/j.mattod.2022.11.02310.1016/j.mattod.2022.11.023 ↗
45. 10.1038/s41586-023-06860-510.1038/s41586-023-06860-5 ↗
46. C. Y.Huang G.Dewey E.Mannebach A.Phan P.Morrow W.Rachmady I. C.Tung N.Thomas U.Alaan R.Paul N.Kabir B.Krist A.Oni M.Mehta M.Harper P.Nguyen R.Keech S.Vishwanath K. L.Cheong J. S.Kang A.Lilak M.Metz S.Clendenning B.Turkot R.Schenker H. J.Yoo M.Radosavljevic J.Kavalieros in2020 IEEE International Electron Devices Meeting (IEDM) San Francisco CA USA2020 pp.20.6.1−20.6.4..
47. 10.1109/5.92964710.1109/5.929647 ↗