Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance
This work provides an off‐chip training method for a one‐selector‐one‐resistor (1S1R) crossbar array (CBA) device with wire resistance (rcc) and nonlinear conductance (g i,j) of 1S1R devices for hardware neural network (HNN) applications. An iterative method is introduced to calculate the node volta...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2022-08-01
|
| Series: | Advanced Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/aisy.202100256 |
| _version_ | 1811339865341558784 |
|---|---|
| author | Jihun Kim Hyo Cheon Woo Sunwoo Lee Byeol Jun Lee Taegyun Park Cheol Seong Hwang |
| author_facet | Jihun Kim Hyo Cheon Woo Sunwoo Lee Byeol Jun Lee Taegyun Park Cheol Seong Hwang |
| author_sort | Jihun Kim |
| collection | DOAJ |
| description | This work provides an off‐chip training method for a one‐selector‐one‐resistor (1S1R) crossbar array (CBA) device with wire resistance (rcc) and nonlinear conductance (g i,j) of 1S1R devices for hardware neural network (HNN) applications. An iterative method is introduced to calculate the node voltages of the 1S1R CBA, which arises from the variable voltage drop through the wires with rcc and g i,j. Several mathematical approximations are introduced for fast and efficient calculation. The proposed method trains the HNN to have an inference accuracy of 85.9%, whereas the inference accuracy of HNN without the rcc consideration drops to 38.5%. The inference running time with the proposed method is 1% of the HSPICE‐based simulation for the given HNN structure. As the rcc increases, the inference accuracy declines due to the decreased device voltage from the target values. The worst voltage model is adopted to identify the design factors that affected the accuracy. A CBA with a size almost three times larger can be used for the HNN if the rcc is appropriately addressed under the given device conditions. |
| first_indexed | 2024-04-13T18:32:55Z |
| format | Article |
| id | doaj.art-19cd9292bb9d478a902e8dc78014ba1d |
| institution | Directory Open Access Journal |
| issn | 2640-4567 |
| language | English |
| last_indexed | 2024-04-13T18:32:55Z |
| publishDate | 2022-08-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advanced Intelligent Systems |
| spelling | doaj.art-19cd9292bb9d478a902e8dc78014ba1d2022-12-22T02:35:00ZengWileyAdvanced Intelligent Systems2640-45672022-08-0148n/an/a10.1002/aisy.202100256Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire ResistanceJihun Kim0Hyo Cheon Woo1Sunwoo Lee2Byeol Jun Lee3Taegyun Park4Cheol Seong Hwang5Department of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaDepartment of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaDepartment of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaDepartment of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaDepartment of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaDepartment of Materials Science and Engineering Seoul National University, and Inter-University Semiconductor Research Center Seoul National University Gwanak-ro 1 Daehag-dong Gwanak-gu Seoul 08826 Republic of KoreaThis work provides an off‐chip training method for a one‐selector‐one‐resistor (1S1R) crossbar array (CBA) device with wire resistance (rcc) and nonlinear conductance (g i,j) of 1S1R devices for hardware neural network (HNN) applications. An iterative method is introduced to calculate the node voltages of the 1S1R CBA, which arises from the variable voltage drop through the wires with rcc and g i,j. Several mathematical approximations are introduced for fast and efficient calculation. The proposed method trains the HNN to have an inference accuracy of 85.9%, whereas the inference accuracy of HNN without the rcc consideration drops to 38.5%. The inference running time with the proposed method is 1% of the HSPICE‐based simulation for the given HNN structure. As the rcc increases, the inference accuracy declines due to the decreased device voltage from the target values. The worst voltage model is adopted to identify the design factors that affected the accuracy. A CBA with a size almost three times larger can be used for the HNN if the rcc is appropriately addressed under the given device conditions.https://doi.org/10.1002/aisy.202100256crossbar arrayIR dropneural networksone-selector-one-resistorselector |
| spellingShingle | Jihun Kim Hyo Cheon Woo Sunwoo Lee Byeol Jun Lee Taegyun Park Cheol Seong Hwang Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance Advanced Intelligent Systems crossbar array IR drop neural networks one-selector-one-resistor selector |
| title | Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance |
| title_full | Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance |
| title_fullStr | Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance |
| title_full_unstemmed | Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance |
| title_short | Training Method for Accurate Off‐Chip Training of One‐Selector‐One‐Resistor Crossbar Array with Nonlinearity and Wire Resistance |
| title_sort | training method for accurate off chip training of one selector one resistor crossbar array with nonlinearity and wire resistance |
| topic | crossbar array IR drop neural networks one-selector-one-resistor selector |
| url | https://doi.org/10.1002/aisy.202100256 |
| work_keys_str_mv | AT jihunkim trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance AT hyocheonwoo trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance AT sunwoolee trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance AT byeoljunlee trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance AT taegyunpark trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance AT cheolseonghwang trainingmethodforaccurateoffchiptrainingofoneselectoroneresistorcrossbararraywithnonlinearityandwireresistance |