MoS₂/Graphene Device for Reconfigurable Fuzzy Logic Hardware” Featured in Nature Electronics
Edge devices face challenges in running deep neural networks (DNNs) due to limited computational resources and power constraints. Fuzzy logic systems, known for their compact size and ability to handle uncertain data, offer a promising alternative for more efficient edge computing. However, building hardware-based fuzzy logic systems remains challenging. Traditional methods for creating reconfigurable membership function generators (MFGs) involve high circuit complexity and power consumption.
In a recent publication in Nature Electronics, researchers from the University of Southern California (USC), the Chinese Academy of Sciences’ Institute of Semiconductors, Northwestern University, and the University of Hong Kong present a new approach.
Led by Hefei Liu, Jiahui Ma, Jiangbin Wu, Mark C. Hersam, and Han Wang, the team developed a multi-gate van der Waals (vdW) interfacial junction transistor (IJT) based on MoS₂/graphene heterostructures. This device can produce adjustable, Gaussian-like and π-shaped membership functions. By integrating these MFGs with peripheral circuits, they created hardware for a reconfigurable fuzzy controller, enabling nonlinear system control.
This fuzzy logic system can also be combined with shallow convolutional neural networks, forming a fuzzy neural network (FNN) with improved image segmentation capabilities.
Key Contributors:
- First Authors: Hefei Liu, Jiangbin Wu, Jiahui Ma
- Corresponding Authors: Mark C. Hersam & Han Wang
- Institutions: Northwestern University, University of Hong Kong
Research Highlights:
- Figure 1: Structure, photoluminescence spectrum, and electrical characteristics of the vdW-IJT device.
- Figure 2: Multi-gate vdW-IJT structure and highly tunable electrical properties.
- Figure 3: Hardware implementation of a fuzzy PID controller.
- Figure 4: Fuzzy neural network (FNN) applied for image segmentation.
- Figure 5: Comparison between reconfigurable MFGs in traditional CMOS and vdW-IJT devices.
For more details, refer to the original paper: Liu, H., Wu, J., Ma, J. et al., A van der Waals Interfacial Junction Transistor for Reconfigurable Fuzzy Logic Hardware, Nature Electronics (2024). DOI: 10.1038/s41928-024-01256-3