Deep Learning to Automate Parameter Extraction and Model Fitting of Two-Dimensional Transistors
Bennett, Uslu, Gault, Khan, Hoang, Peña · cs.LG,cond-mat.mtrl-sci,physics.app-ph · 2026-09-05 · 原文
We present a deep learning approach to extract physical parameters (e.g., mobility, Schottky contact barrier height, defect profiles) of two-dimensional (2D) transistors from electrical measurements, enabling automated parameter extraction and technology computer-aided design (TCAD) fitting. To facilitate this task, we implement a simple data augmentation and pre-training approach by training a secondary neural network to approximate a physics-based device simulator. This method enables high-quality fits after training the neural network on electrical data generated from physics-based simulations of ~500 devices, a factor >40\times fewer than other recent efforts. Consequently, fitting can be achieved by training on physically rigorous TCAD models, including complex geometry, self-consistent transport, and electrostatic effects, and is not limited to computationally inexpensive compact models. We apply our approach to reverse-engineer key parameters from experimental monolayer WS_2 transistors, achieving a median coefficient of determination (R^2) = 0.99 when fitting measured electrical data. We also demonstrate that this approach generalizes and scales well by reverse-engine
讲义
讲义·推断 依据「原文」自动生成的结构化摘要（推断），非原文表述；以原文为准。
1. 人话版
We present a deep learning approach to extract physical parameters (e.g., mobility, Schottky contact barrier height, defect profiles) of two-dimensional (2D) transistors from electrical measurements, enabling automated parameter extraction and technology computer-aided design (TCAD) fitting.
To facilitate this task, we implement a simple data augmentation and pre-training approach by training a secondary neural network to approximate a physics-based device simulator.
2. 领域脉络
本文类目：cs.LG、cond-mat.mtrl-sci、physics.app-ph，属于其所在研究脉络的最新进展。
3. 机制拆解
摘要未展开方法细节——精读时重点看方法/模型部分。
4. 证据与数字
This method enables high-quality fits after training the neural network on electrical data generated from physics-based simulations of ~500 devices, a factor >40\times fewer than other recent efforts.
We apply our approach to reverse-engineer key parameters from experimental monolayer WS_2 transistors, achieving a median coefficient of determination (R^2) = 0.99 when fitting measured electrical data.
5. 反例与边界
Consequently, fitting can be achieved by training on physically rigorous TCAD models, including complex geometry, self-consistent transport, and electrostatic effects, and is not limited to computationally inexpensive compact models.
6. 跨领域连接与意外收获
横跨 3 个类目（cs.LG、cond-mat.mtrl-sci、physics.app-ph），关注其在你兴趣板块间的迁移面。
7. 可复用方法
把本文机制与你手头项目对照，找一个两周内能验证的最小实验。
8. 术语表
精读时把不熟的术语记入此处，作为下次回忆的锚点。