dorsal/arxiv
View SchemaVLM-CAD: VLM-Optimized Collaborative Agent Design Workflow for Analog Circuit Sizing
| Authors | Guanyuan Pan, Yugui Lin, Tiansheng Zhou, Pietro Liò, Shuai Wang, Yaqi Wang |
|---|---|
| Categories | |
| ArXiv ID | 2601.07315vv1 |
| URL | https://arxiv.org/abs/2601.07315 |
| License | http://creativecommons.org/licenses/by-nc-nd/4.0/ |
Abstract
Analog mixed-signal circuit sizing involves complex trade-offs within high-dimensional design spaces. Existing automatic analog circuit sizing approaches often underutilize circuit schematics and lack the explainability required for industry adoption. To tackle these challenges, we propose a Vision Language Model-optimized collaborative agent design workflow (VLM-CAD), which analyzes circuits, optimizes DC operating points, performs inference-based sizing and executes external sizing optimization. We integrate Image2Net to annotate circuit schematics and generate a structured JSON description for precise interpretation by Vision Language Models. Furthermore, we propose an Explainable Trust Region Bayesian Optimization method (ExTuRBO) that employs collaborative warm-starting from agent-generated seeds and offers dual-granularity sensitivity analysis for external sizing optimization, supporting a comprehensive final design report. Experiment results on amplifier sizing tasks using 180nm, 90nm, and 45nm Predictive Technology Models demonstrate that VLM-CAD effectively balances power and performance, achieving a 100% success rate in optimizing an amplifier with a complementary input and a class-AB output stage, while maintaining total runtime under 43 minutes across all experiments.
{
"annotation_id": "dc6ae323-c8c4-46a6-b7e0-9a26596b3b18",
"date_created": "2026-02-17T05:53:12.525000Z",
"date_modified": "2026-02-17T05:53:12.525000Z",
"file_hash": "4989cdfbc9a84182ea76076cabf7e018217e3b3b2340091e6e79305650bc9c63",
"private": false,
"record": {
"abstract": "Analog mixed-signal circuit sizing involves complex trade-offs within high-dimensional design spaces. Existing automatic analog circuit sizing approaches often underutilize circuit schematics and lack the explainability required for industry adoption. To tackle these challenges, we propose a Vision Language Model-optimized collaborative agent design workflow (VLM-CAD), which analyzes circuits, optimizes DC operating points, performs inference-based sizing and executes external sizing optimization. We integrate Image2Net to annotate circuit schematics and generate a structured JSON description for precise interpretation by Vision Language Models. Furthermore, we propose an Explainable Trust Region Bayesian Optimization method (ExTuRBO) that employs collaborative warm-starting from agent-generated seeds and offers dual-granularity sensitivity analysis for external sizing optimization, supporting a comprehensive final design report. Experiment results on amplifier sizing tasks using 180nm, 90nm, and 45nm Predictive Technology Models demonstrate that VLM-CAD effectively balances power and performance, achieving a 100% success rate in optimizing an amplifier with a complementary input and a class-AB output stage, while maintaining total runtime under 43 minutes across all experiments.",
"arxiv_id": "2601.07315",
"authors": [
"Guanyuan Pan",
"Yugui Lin",
"Tiansheng Zhou",
"Pietro Li\u00f2",
"Shuai Wang",
"Yaqi Wang"
],
"categories": [
"cs.MA",
"cs.AI",
"cs.AR"
],
"license": "http://creativecommons.org/licenses/by-nc-nd/4.0/",
"title": "VLM-CAD: VLM-Optimized Collaborative Agent Design Workflow for Analog Circuit Sizing",
"url": "https://arxiv.org/abs/2601.07315",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "321b88ed-265f-4fbc-b91d-c0c5754a2bb0",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}