Skip to main navigation Skip to search Skip to main content

Dual-Loss Trained Compact Cascaded ANN for Custom-Range BSIM-CMG Parameter Extraction

  • Sidhant V. Vazarkar*
  • , Abdullah M.H. Cachhi
  • , Shreyan Ghosh
  • , Girish Pahwa
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper presents a novel deep learning framework that combines dual-loss training with floating normalization for BSIM-CMG model parameter extraction. The framework addresses the fundamental challenge of non-unique parameter-to-characteristic mappings through a cascaded forward-inverse neural network architecture trained with both parameter supervision and curve reconstruction constraints. A flexible floating normalization scheme enables user-defined extraction ranges, allowing multiple valid parameter sets to be extracted from identical I-V characteristics. The dual-loss training mechanism provides superior convergence by simultaneously enforcing parameter accuracy and forward-inverse consistency. Validated with 200 K Monte Carlo samples across 3 nm, 7 nm, and 14 nm FinFET technology nodes, the framework achieves RMSE below 3.75% while offering unprecedented flexibility compared to conventional fixed-range extraction methods.

Original languageEnglish
Title of host publication10th IEEE Electron Devices Technology and Manufacturing Conference
Subtitle of host publicationEmerging Semiconductor Devices and Manufacturing Technologies, EDTM 2026
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331585983
DOIs
StatePublished - 2026
Event10th IEEE Electron Devices Technology and Manufacturing Conference: Emerging Semiconductor Devices and Manufacturing Technologies, EDTM 2026 - Penang, Malaysia
Duration: 1 Mar 20264 Mar 2026

Publication series

Name10th IEEE Electron Devices Technology and Manufacturing Conference: Emerging Semiconductor Devices and Manufacturing Technologies, EDTM 2026

Conference

Conference10th IEEE Electron Devices Technology and Manufacturing Conference: Emerging Semiconductor Devices and Manufacturing Technologies, EDTM 2026
Country/TerritoryMalaysia
CityPenang
Period1/03/264/03/26

Keywords

  • BSIM-CMG
  • deep learning
  • dual-loss training
  • floating normalization
  • inverse design
  • parameter extraction

Fingerprint

Dive into the research topics of 'Dual-Loss Trained Compact Cascaded ANN for Custom-Range BSIM-CMG Parameter Extraction'. Together they form a unique fingerprint.

Cite this