Research map / Probabilistic and scientific modeling / Differential equations and PDEs
LSH-SMILE: Locality Sensitive Hashing Accelerated Simulation and Learning
Chonghao Sima, Yexiang Xue
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 LSH-SMILE instead of brute-force Torch forward simulation | Grain growth simulation · Grain growth simulation | runtime (hr): 16.84 -> 3.86 | Improved |
| Use LSH-SMILE instead of brute-force Torch forward simulation | Nanovoid simulation · Nanovoid simulation | runtime (hr): 53.13 -> 2.13 | Improved |
| Use LSH-SMILE inside backward learning | Grain growth learning · Grain growth learning | runtime (hr): 17.29 -> 4.5 | Improved |
| Use LSH-SMILE for grain growth simulation | Grain growth simulation · Grain growth simulation | memory usage (MB): 1135 -> 870 | Improved |
| Use LSH-SMILE for nanovoid simulation | Nanovoid simulation · Nanovoid simulation | memory usage (MB): 3294 -> 2735 | Improved |
| Use LSH-SMILE for grain growth learning | Grain growth learning · Grain growth learning | memory usage (MB): 1212 -> 889 | Improved |
| Replace exact computation with LSH approximation | Grain growth simulation · Grain growth simulation | difference from ground truth: 10−5 after 180 steps | No clear effect |
| Replace exact computation with LSH approximation | Nanovoid simulation · Nanovoid simulation | difference from ground truth: 10−5 after 100 steps | No clear effect |
| Longer embedded forward simulation in learning (T=30 vs T=10), enabled by LSH-SMILE savings | Grain growth learning · Grain growth learning | learned parameter B (true 1.0): 8.8226 -> 2.01454 | Improved |
| Longer embedded forward simulation in learning (T=30 vs T=10) | Grain growth learning · Grain growth learning | learned kappa (true 0.1): 0.8431 -> 0.0834962 | Improved |
| Longer embedded forward simulation in learning (T=30 vs T=10) | Grain growth learning · Grain growth learning | learned parameter A (true 1.0): 1.9431 -> 1.98483 | Mixed |
| Longer embedded forward simulation in learning (T=30 vs T=10) | Grain growth learning · Grain growth learning | learned parameter L (true 5.0): 9.0462 -> 11.6504 | Hurt |
Closest papers by meaning
- Locality-Sensitive Hashing-Based Efficient Point Transformer with Applications in High-Energy Physics (ICML 2024 · similarity 0.85)
- LSH-MoE: Communication-efficient MoE Training via Locality-Sensitive Hashing (NeurIPS 2024 · similarity 0.85)
- EvoMesh: Adaptive Physical Simulation with Hierarchical Graph Evolutions (ICML 2025 · similarity 0.84)
- Neural Physical Simulation with Multi-Resolution Hash Grid Encoding (AAAI 2024 · similarity 0.84)
- SAFE-SVD: Sensitivity-Aware Fidelity-Enforcing SVD for Physics Foundation Models (NeurIPS 2026 · similarity 0.84)
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs (NeurIPS 2025 · 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.
