Research map / Probabilistic and scientific modeling / Molecules and proteins
BubbleML: A Multiphase Multiphysics Dataset and Benchmarks for Machine Learning
Sheikh Md Shakeel Hassan, Arthur Feeney, Akash Dhruv, Jihoon Kim, Youngjoon Suh, Jaiyoung Ryu, Yoonjin Won, Aparna Chandramowlishwaran
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 |
|---|---|---|---|
| fine-tune RAFT from FlyingChairs on BubbleML (C -> C+B) | RAFT/GMFlow · PB-Saturated | EPE: 4.20 -> 2.33 | Improved |
| fine-tune GMFlow from FlyingChairs on BubbleML (C -> C+B) | RAFT/GMFlow · PB-Saturated | EPE: 4.73 -> 2.36 | Improved |
| fine-tune Sintel-pretrained RAFT on BubbleML (C+T+S -> C+T+S+B) | RAFT/GMFlow · PB-Saturated | EPE: 8.39 -> 2.34 | Improved |
| fine-tune Sintel-pretrained GMFlow on BubbleML (C+T+S -> C+T+S+B) | RAFT/GMFlow · PB-Saturated | EPE: 14.65 -> 2.56 | Improved |
| fine-tune Sintel-pretrained RAFT on BubbleML | RAFT/GMFlow · FB-Gravity | EPE: 21.04 -> 4.12 | Improved |
| fine-tune Sintel-pretrained GMFlow on BubbleML | RAFT/GMFlow · FB-Inlet Velocity | EPE: 47.88 -> 6.88 | Improved |
| fine-tune RAFT on BubbleML (C+T+S -> C+T+S+B) | RAFT/GMFlow · Sintel Clean (Train) | EPE: 0.77 -> 0.89 | Hurt |
| fine-tune RAFT on BubbleML (C+T+S -> C+T+S+B) | RAFT/GMFlow · KITTI F1-EPE (Train) | EPE: 1.54 -> 1.83 | Hurt |
| pool boiling: best starting checkpoint is FlyingChairs (C+B) rather than Sintel (C+T+S+B) | RAFT/GMFlow · PB datasets | EPE: 2.33 vs 2.34 | Mixed |
| flow boiling: starting from Sintel checkpoint (C+T+S+B) beats starting from Chairs (C+B) | RAFT/GMFlow · FB-Inlet Velocity | EPE: 7.01 -> 6.88 | Improved |
| swap RAFT for GMFlow after fine-tuning | RAFT/GMFlow · PB-Saturated | EPE: 2.34 vs 2.56 | No clear effect |
| use UNet-based vs FNO-variant neural PDE solvers for temperature only | neural PDE solvers · BubbleML temperature forecasting | error metrics | Improved |
Closest papers by meaning
- Bubbleformer: Forecasting Boiling with Transformers (NeurIPS 2025 · similarity 0.91)
- Neptuna: A Comprehensive Machine Learning Framework for Benchmarking Complex Multiphase Flows (NeurIPS 2026 · similarity 0.88)
- PINNacle: A Comprehensive Benchmark of Physics-Informed Neural Networks for Solving PDEs (NeurIPS 2024 · similarity 0.84)
- End-to-End Phase Field Model Discovery Combining Experimentation, Crowdsourcing, Simulation and Learning (AAAI 2024 · similarity 0.83)
- The Catechol Benchmark: Time-series Solvent Selection Data for Few-shot Machine Learning (NeurIPS 2025 · similarity 0.83)
- Deep Learning-based Algebraic Reynolds Stress Closures for RANS simulations of Turbulent Flows (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.
