An Interpretable Symbolic Regression Model for Prediction of GaN Vertical Power MOSFET Failsafe Boundaries
- Resource Type
- Conference
- Authors
- Singh, Smriti; Ashai, Aasim; Mukherjee, Ankita; Pramanik, Tanmoy; Sarkar, Biplab
- Source
- 2024 8th IEEE Electron Devices Technology & Manufacturing Conference (EDTM) Electron Devices Technology & Manufacturing Conference (EDTM), 2024 8th IEEE. :1-3 Mar, 2024
- Subject
- Bioengineering
Components, Circuits, Devices and Systems
Computing and Processing
Engineered Materials, Dielectrics and Plasmas
Engineering Profession
Fields, Waves and Electromagnetics
General Topics for Engineers
Photonics and Electrooptics
Semiconductor device modeling
MOSFET
Computational modeling
Voltage
Logic gates
Predictive models
Power transistors
GaN
Vertical transistors
Potential barrier
Genetic algorithms
Symbolic regression
- Language
In this work, a symbolic regression (SR) model has been developed to predict the GaN Fin-MOSFET channel potential profile, particularly at the blocking modes. This model precisely predicts the source-drain barrier which is a pivotal aspect of failsafe device operation. The SR model explains the dependency of different input parameters on the output response of interest. This work is first to report an interpretable approach for quantifying the device design of GaN vertical power transistors for a robust circuit design.