Research map / Graph learning / GNN architectures
Graph of Circuits with GNN for Exploring the Optimal Design Space
Aditya Hemant Shahane, Saripilli Venkata Swapna Manjiri, Ankesh Jain, Sandeep Kumar
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 GraphSAGE instead of GCN in GCX (ablation) | GCX · Two-stage OTA, l=50% | Gain R2: 0.22 -> 0.60 | Improved |
| use GraphSAGE instead of GAT in GCX (ablation) | GCX · Two-stage OTA, l=50% | Gain R2: 0.41 -> 0.60 | Improved |
| use GraphSAGE instead of GCN in GCX (ablation) | GCX · Ring oscillator, l=50% | Frequency R2: 0.44 -> 0.83 | Improved |
| use GraphSAGE instead of GAT in GCX (ablation) | GCX · Ring oscillator, l=50% | Delay R2: 0.58 -> 0.82 | Improved |
| raise labeled fraction from 30% to 50% (ablation) | GCX(SAGE) · Two-stage OTA | Gain R2: 0.42 -> 0.60 | Improved |
| raise labeled fraction from 30% to 50% (ablation) | GCX(SAGE) · Ring oscillator | Frequency R2: 0.45 -> 0.83 | Improved |
| raise labeled fraction from 30% to 50% (ablation) | GCX(SAGE) · Two-stage OTA | Power R2: 0.97 -> 0.99 | Improved |
| replace ESSAB surrogate with GCX(SAGE) | GCX(SAGE) · Two-stage OTA | UGB R2: 0.50 -> 0.60 | Improved |
| replace ESSAB surrogate with GCX(SAGE) | GCX(SAGE) · Ring oscillator | Frequency R2: 0.89 -> 0.83 | Hurt |
| replace ParaGraph graph-level regression with node-level GCX | GCX · Circuit surrogate | R2 | Improved |
| use ASTROG instead of BO-EI | ASTROG · Two-stage OTA | successful runs: 03/10 -> 06/10 | Improved |
| use ASTROG instead of BO-EI | ASTROG · Two-stage OTA | convergence speed: 8x faster | Improved |
Closest papers by meaning
- FALCON: An ML Framework for Fully Automated Layout-Constrained Analog Circuit Design (NeurIPS 2025 · similarity 0.89)
- AnalogToBi: Device-Level Analog Circuit Topology Generation via Bipartite Graph and Grammar Guided Decoding (NeurIPS 2026 · similarity 0.87)
- Circuit Representation Learning with Masked Gate Modeling and Verilog-AIG Alignment (ICLR 2025 · similarity 0.87)
- DeepGate4: Efficient and Effective Representation Learning for Circuit Design at Scale (ICLR 2025 · similarity 0.86)
- AnalogGenie-Lite: Enhancing Scalability and Precision in Circuit Topology Discovery through Lightweight Graph Modeling (ICML 2025 · similarity 0.86)
- CircuitFusion: Multimodal Circuit Representation Learning for Agile Chip Design (ICLR 2025 · similarity 0.86)
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.
