Rapid FinFET Modelling Using an Autoencoder

2026-06-22

AI Summary

This work proposes a compact modeling approach for FinFET devices based on an autoencoder architecture, addressing the complexity and computational cost of conventional modeling techniques that hinder efficient circuit simulation and rapid device characterization. The method explicitly incorporates drain-to-source voltage (VDS) as an input feature for the first time, enabling compression of full I–V curves into a low-dimensional latent space. By leveraging calibrated BSIM-CMG data, the model accurately reconstructs I–V characteristics with high fidelity using only a limited amount of training data. Furthermore, key device parameters—including threshold voltage (VTH), subthreshold swing (SS), and peak transconductance (gm)—can be directly and precisely extracted from the latent representation, significantly enhancing both the efficiency and accuracy of FinFET characterization and simulation.

Abstract

This work presents a machine learning framework that leverages an autoencoder (AE) for the efficient modeling of FinFET. We first calibrated a BSIM-CMG model to generate a dataset of current-voltage (ID-VG) characteristics. This data was used to train an autoencoder that compresses full I-V curves into a low-dimensional latent space, which intrinsically encodes key device physics. A key innovation is the explicit incorporation of parameter such as drain to source voltage (VDS) as an input feature, enhancing the model ability to capture bias dependent variation. The trained model successfully reconstructs full I-V curves and directly extracts critical device metrics including threshold voltage (VTH), subthreshold slope (SS), and peak transconductance (gm). This approach demonstrates that data driven compact models, built from actual characterization data, can achieve high accuracy with minimal training data, providing a powerful tool for rapid device characterization, modelling and circuit level simulation.

Problem

• FinFET modeling
• autoencoder
• I-V characteristics
• compact model
• device characterization

Innovation

• Autoencoder
• FinFET modeling
• Compact modeling
• Latent space representation
• Bias-dependent characterization