📢 Active Learning for Neural PDE Solvers (AL4PDE; arxiv.org/abs/2408.01536) While active learning is common in other domains, it has yet to be studied extensively for neural PDE solving. This work introduces AL4PDE, a modular and extensible active learning benchmark. #ML4PDE
Includes multiple parametric PDEs & SOTA surrogate models for a solver-in-the-loop setting, enabling the evaluation of existing & development of new active learning methods for PDE solving. Evaluates batch active learning algorithms such as uncertainty & feature-based methods.
Shows that active learning reduces the average error by up to 71% compared to random sampling & significantly reduces worst-case errors. Acquired datasets are reusable, providing benefits for surrogate models not involved in the data generation. Code and dataset available soon.
A new line of research towards developing algorithms & methods that enable neural operators to achieve generalization over desiderata such as PDE parameters, ICs, BCs, Geometries, Resolutions, etc. #NeuralOperators #ML4Science #PDEs #NumericalSolvers #ActiveLearning #AI4Science
United States เทรนด์
- 1. #WorldSeries 87.7K posts
- 2. Blue Jays 56.8K posts
- 3. Luka 52.4K posts
- 4. Snell 12.6K posts
- 5. #SmackDown 29K posts
- 6. #Dodgers 13.1K posts
- 7. #BostonBlue 4,682 posts
- 8. Addison Barger 8,861 posts
- 9. Sheehan 1,890 posts
- 10. Paolo 13K posts
- 11. Grand Slam 15.8K posts
- 12. #WANTITALL 31.2K posts
- 13. Knicks 26.2K posts
- 14. Celtics 21.5K posts
- 15. Zion 20.2K posts
- 16. Kyshawn George 1,608 posts
- 17. Darryn Peterson 2,419 posts
- 18. Halo 147K posts
- 19. Cole Anthony 1,783 posts
- 20. Paul Reed N/A
Something went wrong.
Something went wrong.