Boosting High-Speed Channel Transformer Performance With Gating Neural Network

Author

• Davit Kharshiladze
• Hyunwook Park
• Zhiping Yang, Missouri University of Science and TechnologyFollow
• Chulsoon Hwang, Missouri University of Science and TechnologyFollow

Abstract

This paper proposes a method to improve the performance of high-speed channel transformer (HSCT) neural network, by using an additional gating network. HSCT is a transformer-based neural network, capable of predicting RLGC and S-Parameters of striplines, based on cross-sectional definition. The performance boost is achieved by choosing between two pre-trained HSCTs to determine an optimal outcome. Each HSCT is trained on different output data normalization, emphasizing varied behavior of the model.

Recommended Citation

D. Kharshiladze et al., "Boosting High-Speed Channel Transformer Performance With Gating Neural Network," 30th IEEE Workshop on Signal and Power Integrity Spi 2026, Institute of Electrical and Electronics Engineers, Jan 2026.

The definitive version is available at https://doi.org/10.1109/SPI68887.2026.11594680

Department(s)

Electrical and Computer Engineering

Keywords and Phrases

data normalization; gating neural network; High-speed channel; mixture of experts

Document Type

Article - Conference proceedings

Document Version

Citation

File Type

text

Language(s)

English

Rights

© 2026 Institute of Electrical and Electronics Engineers, All rights reserved.

Publication Date

01 Jan 2026