MOTIF-RF: Multi-template On-chip Transformer Synthesis Incorporating Frequency-domain Self-transfer Learning for RFIC Design Automation

2025-11-26

AI Summary

Reverse design and accurate modeling of on-chip transformers (XFMRs) in radio-frequency integrated circuits (RFICs) remain challenging due to complex electromagnetic behavior and limited generalizability across frequency bands. Method: This paper proposes a synergistic framework integrating multi-template deep learning surrogate modeling with frequency-domain self-transfer learning. It systematically evaluates four surrogate architectures—MLP, CNN, U-Net, and graph transformer—for S-parameter prediction; introduces a novel frequency-domain self-transfer learning strategy leveraging inter-band spectral correlation to enhance cross-band generalization; and couples the surrogate with CMA-ES for end-to-end inverse design. Contribution/Results: The framework reduces S-parameter prediction error by 30–50% across multi-impedance matching tasks, achieves rapid design convergence and high solution fidelity, and—critically—enables, for the first time, a fully AI-driven, automated workflow from electrical specifications to GDSII layout, significantly advancing the practical deployment of RFIC design automation.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Search and Optimization: Metareasoning and Metaheuristics
• Planning, Routing, and Scheduling: Planning/Scheduling and Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

This paper presents a systematic study on developing multi-template machine learning (ML) surrogate models and applying them to the inverse design of transformers (XFMRs) in radio-frequency integrated circuits (RFICs). Our study starts with benchmarking four widely used ML architectures, including MLP-, CNN-, UNet-, and GT-based models, using the same datasets across different XFMR topologies. To improve modeling accuracy beyond these baselines, we then propose a new frequency-domain self-transfer learning technique that exploits correlations between adjacent frequency bands, leading to around 30%-50% accuracy improvement in the S-parameters prediction. Building on these models, we further develop an inverse design framework based on the covariance matrix adaptation evolutionary strategy (CMA-ES) algorithm. This framework is validated using multiple impedance-matching tasks, all demonstrating fast convergence and trustworthy performance. These results advance the goal of AI-assisted specs-to-GDS automation for RFICs and provide RFIC designers with actionable tools for integrating AI into their workflows.

Problem

• Develops multi-template ML surrogate models for RFIC transformer design
• Proposes frequency-domain self-transfer learning to enhance S-parameters prediction accuracy
• Creates an inverse design framework using CMA-ES for impedance-matching tasks

Innovation

• Multi-template machine learning surrogate models for RFIC design
• Frequency-domain self-transfer learning improves S-parameters prediction
• Inverse design framework using CMA-ES algorithm for fast convergence