Researchers have developed CADET, a system for generating synthetic printed circuit board (PCB) routing instances along with demonstrations of the electronic design automation (EDA) engine actions. The resulting RouteDemos dataset contains 100,000 boards routed without design rule violations, mimicking real-world board distributions. Training a routing policy using behavior cloning and reinforcement learning on this dataset improved the rate of complete, violation-free routing from 38.0% to 67.2% on real test boards, outperforming baselines including GPT-6 Astra. AI
IMPACT This research could lead to more efficient and automated PCB design processes, potentially reducing development time and costs in electronics manufacturing.
RANK_REASON The cluster contains an academic paper detailing a new dataset and methodology for AI-driven PCB routing. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- CADET
- FreeRouting
- GPT-6
- imitation learning
- printed circuit board
- reinforcement learning
- RouteDemos
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →