| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Functional Nanomaterials 2024
|
| Presentation Title |
Photoactive Nanocomposites-emulated Neuromorphic Intelligence |
| Author(s) |
Jae Gwang Kim, Ruochen Liu, Jingjing(Jenny) Qiu, Shiren Wang |
| On-Site Speaker (Planned) |
Shiren Wang |
| Abstract Scope |
ZnO nanoparticles-filled polyvinylpyrrolidone composite is designed to emulate the brain associative learning behavior through stimulus-responsive resistivity switching. Photoactive ZnO nanoparticles were used to emulate the volume transmission mechanisms, and diffused Ag+ ions were to simulate the wire transmission in neuroscience. The artificial synapses of Ag/ZnO-PVP/ITO were further assembled to a 5x5 crossbar, demonstrating fast (1 s) and reliable memory operation with a long-time(1-day) association of electrical and optical signals. This cross-bar array was further to emulate neuromorphic associative computing and image recognition, showing a high accuracy of 92%. Even in the data-sparse condition, image recognition accuracy was still as high as 82%, much higher than that of the conventional artificial neural networks(71%). The emulation of the brain neuromorphic learning process through hierarchical materials and structures suggests a new way for size, weight, and power (SWaP)-efficient neuromorphic intelligence. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Nanotechnology, Composites, Machine Learning |