New benchmark suite ParasGB released for circuit graph learning

Researchers have introduced ParasGB, a new open-source benchmark suite designed for predicting parasitic parameters in analog and mixed-signal (AMS) circuits during the pre-layout phase. This suite addresses the lack of high-fidelity benchmarks for reproducible evaluation of graph neural network (GNN) based parasitic modeling. ParasGB includes large-scale, heterogeneous RC networks extracted from proven designs, along with a unified evaluation protocol and standardized training pipeline for diverse GNN architectures. The benchmark aims to facilitate reproducible research in circuit graph learning and parasitic-aware model development. AI

IMPACT Establishes a standardized benchmark for parasitic estimation in circuit design, potentially accelerating the development of AI models for electronic design automation.

RANK_REASON The cluster contains a research paper introducing a new benchmark suite for a specific machine learning application. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• AMS circuits
• arXiv
• graph neural network
• Hugging Face
• ParasGB
• RC networks

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →