Research map / Graph learning / Graph self-supervised learning
Learning Semantic Representations to Verify Hardware Designs
Shobha Vasudevan, Wenjie Jiang, David Bieber, Rishabh Singh, HAMID SHOJAEI, C. Richard Ho, Charles Sutton
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 |
|---|---|---|---|
| add cover-point graph structure via node sequence embeddings (vs MLP on one-hot cover point id) | coverage prediction baselines · IBEX v1/v2/TPU | accuracy: 56.8 -> 63.2 (IBEX v1, 90%); 34.7 -> 88.6 (TPU, 90%) | Improved |
| replace shallow node sequence embedding with GNN-based Design2Vec (RTL IPA-GNN) | coverage prediction · IBEX v1/v2/TPU | accuracy: 63.2 -> 77.8 (IBEX v1, 90%); 88.6 -> 91.1 (TPU, 90%) | Improved |
| use RTL IPA-GNN instead of GCN/GGNN/GNN-MLP as GNN layer (ablation) | Design2Vec · IBEX v2 | accuracy: 80.3 vs 73.9 (GCN), 76.0 (GGNN), 71.2 (GNN-MLP) at 90% | Improved |
| use RTL IPA-GNN instead of other GNN layers (ablation) | Design2Vec · TPU | accuracy: 91.1 vs 91.1 (GCN) and 91.2 (GGNN) at 90% | No clear effect |
| LSTM instead of mean aggregation of cover point node embeddings (ablation) | Design2Vec RTL IPA-GNN · IBEX v2 | accuracy: 78.9 -> 80.3 (90%) | Improved |
| Design2Vec-guided test generation vs Vizier SRP black-box optimizer for hard cover points | test generation · IBEX v1 hard points | simulator calls to cover point: 93 -> 25 (point 886); 24 -> 3 (point 882) | Improved |
| Design2Vec test generation vs Vizier on TPU hard points | test generation · TPU hard points | simulator calls to cover point: 90 -> 2 (point 36996) | Improved |
Closest papers by meaning
- CircuitFusion: Multimodal Circuit Representation Learning for Agile Chip Design (ICLR 2025 · similarity 0.86)
- Retrieval-Guided Reinforcement Learning for Boolean Circuit Minimization (ICLR 2024 · similarity 0.85)
- Bridging Layout and RTL: Knowledge Distillation based Timing Prediction (ICML 2025 · similarity 0.85)
- Neural Model Checking (NeurIPS 2024 · similarity 0.85)
- DeepLayout: Learning Neural Representations of Circuit Placement Layout (ICML 2025 · similarity 0.84)
- CircuitSense: A Hierarchical Circuit System Benchmark Bridging Visual Comprehension and Symbolic Reasoning in Engineering Design Process (ICLR 2026 · similarity 0.84)
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.
