Emulating synapses for brain-inspired computing by defect engineering in oxide thin films
Memristors are promising for neuromorphic computing, due to its low energy consumption and learning behaviour that mimic the synapses in the neural networks. The uprising of artificial intelligence (AI) has greatly increased the requirements for performance of a computing system. Computing systems a...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/165737 |
| _version_ | 1811696794164264960 |
|---|---|
| author | Wong, Wei Jie |
| author2 | Yu Jing |
| author_facet | Yu Jing Wong, Wei Jie |
| author_sort | Wong, Wei Jie |
| collection | NTU |
| description | Memristors are promising for neuromorphic computing, due to its low energy consumption and learning behaviour that mimic the synapses in the neural networks. The uprising of artificial intelligence (AI) has greatly increased the requirements for performance of a computing system. Computing systems are required to be fast, larger in storage capacity and low energy consumption in order to allow AI programs to operate efficiently and with low energy consumption. Neuromorphic computing has been one of the suitable candidates that is used for running AI systems due to its storage size and processing speed, and special features where processors and memories are located in the same devices as compared to traditional von Neumann architecture where processors and memories are separated.
This project involves the synthesis and electrical testing of NaNbO3 (NNO) oxide thin films for neuromorphic computing to mimic synaptic learning behaviour in human brain. The main purpose is to evaluate the performance of NNO thin films as a memristor for neuromorphic computing since there has not been any research studies that focus on this material for neuromorphic computing.
This study used a furnace for target preparation, a RF sputtering machine for the deposition of NNO thin films and a probe station to characterize the electric behaviour of memristor devices. The NNO thin films have been demonstrated to achieve the synaptic learning behaviour that is similar to the human brain such as Long Term Potentiation (LTP), Long Term Depression (LTD), Short Term Potentiation (STP) and Short Term Depression (STD) synaptic pulses.
To conclude, NNO thin films are promising candidates as a synaptic electronics neuromorphic computing. Further studies can be performed to further optimize the thin film growth and improve the design of memristor devices. In the long term, artificial neural networks based on NNO memristor can be built to run through a series of algorithm such as ANNs and SNNs to realize real AI applications such as the image recognition and large-scale language models such as ChatGPT. |
| first_indexed | 2024-10-01T07:45:01Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/165737 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T07:45:01Z |
| publishDate | 2023 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1657372023-04-15T16:46:05Z Emulating synapses for brain-inspired computing by defect engineering in oxide thin films Wong, Wei Jie Yu Jing School of Materials Science and Engineering Agency for Science, Technology and Research (A*STAR) A*STAR Institute of Material Research and Engineering Liu Huajun yujing@ntu.edu.sg Engineering::Materials Memristors are promising for neuromorphic computing, due to its low energy consumption and learning behaviour that mimic the synapses in the neural networks. The uprising of artificial intelligence (AI) has greatly increased the requirements for performance of a computing system. Computing systems are required to be fast, larger in storage capacity and low energy consumption in order to allow AI programs to operate efficiently and with low energy consumption. Neuromorphic computing has been one of the suitable candidates that is used for running AI systems due to its storage size and processing speed, and special features where processors and memories are located in the same devices as compared to traditional von Neumann architecture where processors and memories are separated. This project involves the synthesis and electrical testing of NaNbO3 (NNO) oxide thin films for neuromorphic computing to mimic synaptic learning behaviour in human brain. The main purpose is to evaluate the performance of NNO thin films as a memristor for neuromorphic computing since there has not been any research studies that focus on this material for neuromorphic computing. This study used a furnace for target preparation, a RF sputtering machine for the deposition of NNO thin films and a probe station to characterize the electric behaviour of memristor devices. The NNO thin films have been demonstrated to achieve the synaptic learning behaviour that is similar to the human brain such as Long Term Potentiation (LTP), Long Term Depression (LTD), Short Term Potentiation (STP) and Short Term Depression (STD) synaptic pulses. To conclude, NNO thin films are promising candidates as a synaptic electronics neuromorphic computing. Further studies can be performed to further optimize the thin film growth and improve the design of memristor devices. In the long term, artificial neural networks based on NNO memristor can be built to run through a series of algorithm such as ANNs and SNNs to realize real AI applications such as the image recognition and large-scale language models such as ChatGPT. Bachelor of Engineering (Materials Engineering) 2023-04-10T09:22:45Z 2023-04-10T09:22:45Z 2023 Final Year Project (FYP) Wong, W. J. (2023). Emulating synapses for brain-inspired computing by defect engineering in oxide thin films. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/165737 https://hdl.handle.net/10356/165737 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Materials Wong, Wei Jie Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title | Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title_full | Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title_fullStr | Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title_full_unstemmed | Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title_short | Emulating synapses for brain-inspired computing by defect engineering in oxide thin films |
| title_sort | emulating synapses for brain inspired computing by defect engineering in oxide thin films |
| topic | Engineering::Materials |
| url | https://hdl.handle.net/10356/165737 |
| work_keys_str_mv | AT wongweijie emulatingsynapsesforbraininspiredcomputingbydefectengineeringinoxidethinfilms |