Research map / Graph learning / GNN architectures
CktGNN: Circuit Graph Neural Network for Electronic Design Automation
Zehao Dong, Weidong Cao, Muhan Zhang, Dacheng Tao, Yixin Chen, Xuan Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use CktGNN instead of PACE | Circuit-property prediction on OCB · OCB | Gain RMSE: 0.644 ± 0.003 -> 0.607 ± 0.003 | Improved |
| Use CktGNN instead of D-VAE | Circuit-property prediction on OCB · OCB | FoM RMSE: 0.897 ± 0.003 -> 0.854 ± 0.003 | Improved |
| Use CktGNN instead of D-VAE | Circuit-property prediction on OCB · OCB | PM Pearson's r: 0.301 ± 0.002 -> 0.217 ± 0.001 | Hurt |
| Use CktGNN instead of PACE | Circuit topology reconstruction · OCB | Topology reconstruction accuracy: 0.306 -> 0.397 | Improved |
| Use CktGNN instead of PACE | Circuit generation · OCB | Valid DAGs (%): 83.12 -> 98.92 | Improved |
| Use CktGNN instead of D-VAE | Circuit generation · OCB | Valid circuits (%): 73.93 -> 98.92 | Improved |
| Use CktGNN instead of D-VAE | Circuit generation · OCB | Novel circuits (%): 97.15 -> 92.29 | Hurt |
| Use CktGNN instead of PACE | Bayesian optimization of circuit designs · OCB | BO FoM: 33.274162 -> 33.436447 | Improved |
| Use CktGNN instead of DAGNN | Circuit-property prediction on Ckt-Bench101 · Ckt-Bench101 | Gain RMSE: 0.409 ± 0.002 -> 0.682 ± 0.004 | Hurt |
| Use CktGNN instead of DAGNN | Circuit-property prediction on Ckt-Bench101 · Ckt-Bench101 | FoM RMSE: 0.639 ± 0.002 -> 0.597 ± 0.009 | Improved |
| Use CktGNN instead of DAGNN | Bayesian optimization with Bohamiann · Ckt-Bench301 | Best FoM detected: 168.61 ± 11.25 -> 175.85 ± 11.26 | Improved |
| Use CktGNN instead of DAGNN | Bayesian optimization with DNGO · Ckt-Bench301 | Best FoM detected: 162.58 ± 15.33 -> 163.29± 16.57 | Mixed |
Later papers that compared with this method
2 later accepted papers report CktGNN as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- AnalogGenie-Lite: Enhancing Scalability and Precision in Circuit Topology Discovery through Lightweight Graph Modeling (ICML 2025 · better on 4 of 4 shared results)
- Directed Graph Grammars for Sequence-based Learning (ICML 2025 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PACE — PACE: A Parallelizable Computation Encoder for Directed Acyclic Graphs (ICML 2022) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FALCON: An ML Framework for Fully Automated Layout-Constrained Analog Circuit Design (NeurIPS 2025 · similarity 0.90)
- AnalogToBi: Device-Level Analog Circuit Topology Generation via Bipartite Graph and Grammar Guided Decoding (NeurIPS 2026 · similarity 0.90)
- AnalogGenie: A Generative Engine for Automatic Discovery of Analog Circuit Topologies (ICLR 2025 · similarity 0.89)
- Circuit Representation Learning with Masked Gate Modeling and Verilog-AIG Alignment (ICLR 2025 · similarity 0.89)
- LaMAGIC: Language-Model-based Topology Generation for Analog Integrated Circuits (ICML 2024 · similarity 0.88)
- DeepGate4: Efficient and Effective Representation Learning for Circuit Design at Scale (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
