CMOS compatible building blocks
This project explores the application of physical reservoir computing in predicting alcohol sales and highlights its potential to match the recurrent neural networks in performance with less complexity and improve efficiency when dealing with sequential data taskings. The project further explores th...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
Nanyang Technological University
2024
|
| 主題: | |
| 在線閱讀: | https://hdl.handle.net/10356/176832 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | This project explores the application of physical reservoir computing in predicting alcohol sales and highlights its potential to match the recurrent neural networks in performance with less complexity and improve efficiency when dealing with sequential data taskings. The project further explores the fine-tuning of the transistor model parameters and demonstrates how a physical system with non-linearity can be harnessed to reduce training load and latency. PRC models leverage the dynamics of a physical system such as transistor model to enable diverse applications for a rapid and energy-efficient output. The results demonstrate that PRC delivers better results to RNN for sequential data tasking and reveals the potential of transistor base model as an alternative to RNN while enhancing computational efficiency and speed. This is a prospective impact on advancing artificial intelligence and machine learning models. |
|---|