CircuitGate advances AI for logic synthesis with functional modeling

Researchers have developed CircuitGate, a novel framework for learning representations of And-Inverter Graphs (AIGs) that goes beyond traditional gate-level analysis. This new approach focuses on circuit-level functional modeling by explicitly encoding global primary-input support and modeling reconvergence between fanins, incorporating logic-inspired Boolean constraints for functional consistency. Evaluations on benchmarks like ForgeEDA, EPFL, and ITC'99 show CircuitGate outperforming existing methods in tasks such as equivalent-gate identification and signal-probability prediction, demonstrating its effectiveness in capturing circuit-level functional dependencies. AI

IMPACT This research could lead to more efficient and robust digital system design by improving the accuracy of AI models in understanding circuit functionality.

RANK_REASON The cluster contains a research paper detailing a new framework for circuit-level functional modeling in electronic design automation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• And-Inverter Graphs
• CircuitGate
• electronic design automation
• ForgeEDA
• graph neural networks
• ITC'99
• OpenABC: An Open-Source Active Breathing Control System for Low-Resource Centers
• Swiss Federal Institute of Technology in Lausanne

• paper
• infra

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