Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update
Among many artificial neural networks, the research on Spike Neural Network (SNN), which mimics the energy-efficient signal system in the brain, is drawing much attention. Memristor is a promising candidate as a synaptic component for hardware implementation of SNN, but several non-ideal device prop...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-03-01
|
| Series: | Frontiers in Computational Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fncom.2021.646125/full |
| _version_ | 1818575946836344832 |
|---|---|
| author | Taeyoon Kim Suman Hu Jaewook Kim Joon Young Kwak Jongkil Park Suyoun Lee Inho Kim Jong-Keuk Park YeonJoo Jeong |
| author_facet | Taeyoon Kim Suman Hu Jaewook Kim Joon Young Kwak Jongkil Park Suyoun Lee Inho Kim Jong-Keuk Park YeonJoo Jeong |
| author_sort | Taeyoon Kim |
| collection | DOAJ |
| description | Among many artificial neural networks, the research on Spike Neural Network (SNN), which mimics the energy-efficient signal system in the brain, is drawing much attention. Memristor is a promising candidate as a synaptic component for hardware implementation of SNN, but several non-ideal device properties are making it challengeable. In this work, we conducted an SNN simulation by adding a device model with a non-linear weight update to test the impact on SNN performance. We found that SNN has a strong tolerance for the device non-linearity and the network can keep the accuracy high if a device meets one of the two conditions: 1. symmetric LTP and LTD curves and 2. positive non-linearity factors for both LTP and LTD. The reason was analyzed in terms of the balance between network parameters as well as the variability of weight. The results are considered to be a piece of useful prior information for the future implementation of emerging device-based neuromorphic hardware. |
| first_indexed | 2024-12-15T00:47:48Z |
| format | Article |
| id | doaj.art-8205df68ea0c40e88455dee99edc8d82 |
| institution | Directory Open Access Journal |
| issn | 1662-5188 |
| language | English |
| last_indexed | 2024-12-15T00:47:48Z |
| publishDate | 2021-03-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Computational Neuroscience |
| spelling | doaj.art-8205df68ea0c40e88455dee99edc8d822022-12-21T22:41:30ZengFrontiers Media S.A.Frontiers in Computational Neuroscience1662-51882021-03-011510.3389/fncom.2021.646125646125Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight UpdateTaeyoon KimSuman HuJaewook KimJoon Young KwakJongkil ParkSuyoun LeeInho KimJong-Keuk ParkYeonJoo JeongAmong many artificial neural networks, the research on Spike Neural Network (SNN), which mimics the energy-efficient signal system in the brain, is drawing much attention. Memristor is a promising candidate as a synaptic component for hardware implementation of SNN, but several non-ideal device properties are making it challengeable. In this work, we conducted an SNN simulation by adding a device model with a non-linear weight update to test the impact on SNN performance. We found that SNN has a strong tolerance for the device non-linearity and the network can keep the accuracy high if a device meets one of the two conditions: 1. symmetric LTP and LTD curves and 2. positive non-linearity factors for both LTP and LTD. The reason was analyzed in terms of the balance between network parameters as well as the variability of weight. The results are considered to be a piece of useful prior information for the future implementation of emerging device-based neuromorphic hardware.https://www.frontiersin.org/articles/10.3389/fncom.2021.646125/fullspiking neural networkmemristornon-linearityhomeostasisLTP/LTD ratio |
| spellingShingle | Taeyoon Kim Suman Hu Jaewook Kim Joon Young Kwak Jongkil Park Suyoun Lee Inho Kim Jong-Keuk Park YeonJoo Jeong Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update Frontiers in Computational Neuroscience spiking neural network memristor non-linearity homeostasis LTP/LTD ratio |
| title | Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update |
| title_full | Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update |
| title_fullStr | Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update |
| title_full_unstemmed | Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update |
| title_short | Spiking Neural Network (SNN) With Memristor Synapses Having Non-linear Weight Update |
| title_sort | spiking neural network snn with memristor synapses having non linear weight update |
| topic | spiking neural network memristor non-linearity homeostasis LTP/LTD ratio |
| url | https://www.frontiersin.org/articles/10.3389/fncom.2021.646125/full |
| work_keys_str_mv | AT taeyoonkim spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT sumanhu spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT jaewookkim spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT joonyoungkwak spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT jongkilpark spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT suyounlee spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT inhokim spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT jongkeukpark spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate AT yeonjoojeong spikingneuralnetworksnnwithmemristorsynapseshavingnonlinearweightupdate |