Research map / Image generation / Autoregressive and latent generation
HubRouter: Learning Global Routing via Hub Generation and Pin-hub Connection
Xingbo Du, Chonghua Wang, Ruizhe Zhong, Junchi Yan
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 routers as extra input to hub generation (hubs -> hubs + routers) (ablation) | HubRouter (GAN) · ISPD-07 | WLR: 1.031 -> 1.022 | Improved |
| add stripe masks to hub generation input (ablation) | HubRouter (GAN) · ISPD-07 | WLR: 1.022 -> 1.005 | Improved |
| use 8 test-time rotation/swap transformations (ablation) | HubRouter (GAN) · ISPD-07 | WLR: 1.087 -> 1.005 | Improved |
| replace R-MST connection model with actor-critic RSMT (T=8) (ablation) | HubRouter (GAN) · ISPD-07 | WLR: 1.099 -> 1.005 | Improved |
| two-phase hub generation and connection instead of direct route generation | HubRouter vs PRNet · ISPD-07 Route-large | correctness rate (%): 0.040 -> 1.000 | Improved |
| hub generation instead of end-to-end route generation, GAN variant | HubRouter (GAN) vs PRNet · IBM01 wirelength | WL: 61950 -> 61056 | Improved |
| hub generation vs PRNet | HubRouter (GAN) vs PRNet · IBM06 wirelength | WL: 287777 -> 278198 | Improved |
| hub generation vs PRNet, total inference time | HubRouter (VAE) vs PRNet · IBM06 time | seconds: 1227.31 -> 17.28 | Improved |
| use VAE instead of GAN as generative model (ablation) | HubRouter · IBM01 wirelength | WL: 61056 -> 64703 | Hurt |
| use DPM instead of GAN as generative model (ablation) | HubRouter · ISPD-07 Route-large generation time | GPU sec: 9.75 -> 672.86 | Hurt |
| use DPM instead of GAN as generative model (ablation) | HubRouter · GRL-16 wirelength | WL: 7768 -> 7746 | Improved |
| use HubRouter instead of REST for RSMT construction at degree 50 | HubRouter · RSMT random points | average percent error: 0.36 -> 0.06 | Improved |
Closest papers by meaning
- NeuralSteiner: Learning Steiner Tree for Overflow-avoiding Global Routing in Chip Design (NeurIPS 2024 · similarity 0.89)
- DSBRouter: End-to-end Global Routing via Diffusion Schr\"{o}dinger Bridge (ICML 2025 · similarity 0.87)
- PreRoutGNN for Timing Prediction with Order Preserving Partition: Global Circuit Pre-training, Local Delay Learning and Attentional Cell Modeling (AAAI 2024 · similarity 0.84)
- Learning to Solve Compositional Geometry Routing Problems (NeurIPS 2026 · similarity 0.84)
- VL-RouterBench: A Benchmark for Vision-Language Model Routing (CVPR 2026 · similarity 0.83)
- SpecDrop: Parameter-Free Category-Conditioned Routing for Modular Specialization (NeurIPS 2026 · similarity 0.83)
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.
