Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse
Abstract Neural reuse can drive organisms to generalize knowledge across various tasks during learning. However, existing devices mostly focus on architectures rather than network functions, lacking the mimic capabilities of neural reuse. Here, we demonstrate a rational device designed based on ferr...
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Nature Communications |
| Online Access: | https://doi.org/10.1038/s41467-024-55701-0 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832594576628514816 |
|---|---|
| author | Jiangang Chen Zhixing Wen Fan Yang Renji Bian Qirui Zhang Er Pan Yuelei Zeng Xiao Luo Qing Liu Liang-Jian Deng Fucai Liu |
| author_facet | Jiangang Chen Zhixing Wen Fan Yang Renji Bian Qirui Zhang Er Pan Yuelei Zeng Xiao Luo Qing Liu Liang-Jian Deng Fucai Liu |
| author_sort | Jiangang Chen |
| collection | DOAJ |
| description | Abstract Neural reuse can drive organisms to generalize knowledge across various tasks during learning. However, existing devices mostly focus on architectures rather than network functions, lacking the mimic capabilities of neural reuse. Here, we demonstrate a rational device designed based on ferroionic CuInP2S6, to accomplish the neural reuse function, enabled by dynamic allocation of the ferro-ionic phase. It allows for dynamic refresh and collaborative work between volatile and non-volatile modes to support the entire neural reuse process. Notably, ferroelectric polarization can remain consistent even after undergoing the refresh process, providing a foundation for the shared functionality across multiple tasks. By implementing neural reuse, the classification accuracy of neuromorphic hardware can improve by 17%, while the consumption is reduced by 40%; in multi-task scenarios, its training speed is accelerated by 2200%, while its generalization ability is enhanced by 21%. Our results are promising towards building refreshable hardware platforms based on ferroelectric-ionic combination capable of accommodating more efficient algorithms and architectures. |
| format | Article |
| id | doaj-art-cb369ca973734d0d8b7560058c181d99 |
| institution | Kabale University |
| issn | 2041-1723 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Nature Communications |
| spelling | doaj-art-cb369ca973734d0d8b7560058c181d992025-01-19T12:30:52ZengNature PortfolioNature Communications2041-17232025-01-011611910.1038/s41467-024-55701-0Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuseJiangang Chen0Zhixing Wen1Fan Yang2Renji Bian3Qirui Zhang4Er Pan5Yuelei Zeng6Xiao Luo7Qing Liu8Liang-Jian Deng9Fucai Liu10School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaSchool of Mathematical Sciences, University of Electronic Science and Technology of ChinaSchool of Optoelectronic Science and Engineering, University of Electronic Science and Technology of ChinaAbstract Neural reuse can drive organisms to generalize knowledge across various tasks during learning. However, existing devices mostly focus on architectures rather than network functions, lacking the mimic capabilities of neural reuse. Here, we demonstrate a rational device designed based on ferroionic CuInP2S6, to accomplish the neural reuse function, enabled by dynamic allocation of the ferro-ionic phase. It allows for dynamic refresh and collaborative work between volatile and non-volatile modes to support the entire neural reuse process. Notably, ferroelectric polarization can remain consistent even after undergoing the refresh process, providing a foundation for the shared functionality across multiple tasks. By implementing neural reuse, the classification accuracy of neuromorphic hardware can improve by 17%, while the consumption is reduced by 40%; in multi-task scenarios, its training speed is accelerated by 2200%, while its generalization ability is enhanced by 21%. Our results are promising towards building refreshable hardware platforms based on ferroelectric-ionic combination capable of accommodating more efficient algorithms and architectures.https://doi.org/10.1038/s41467-024-55701-0 |
| spellingShingle | Jiangang Chen Zhixing Wen Fan Yang Renji Bian Qirui Zhang Er Pan Yuelei Zeng Xiao Luo Qing Liu Liang-Jian Deng Fucai Liu Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse Nature Communications |
| title | Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse |
| title_full | Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse |
| title_fullStr | Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse |
| title_full_unstemmed | Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse |
| title_short | Refreshable memristor via dynamic allocation of ferro-ionic phase for neural reuse |
| title_sort | refreshable memristor via dynamic allocation of ferro ionic phase for neural reuse |
| url | https://doi.org/10.1038/s41467-024-55701-0 |
| work_keys_str_mv | AT jiangangchen refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT zhixingwen refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT fanyang refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT renjibian refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT qiruizhang refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT erpan refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT yueleizeng refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT xiaoluo refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT qingliu refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT liangjiandeng refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse AT fucailiu refreshablememristorviadynamicallocationofferroionicphaseforneuralreuse |