Releases: SciML/Surrogates.jl
Releases · SciML/Surrogates.jl
v6.5.0
Surrogates v6.5.0
Closed issues:
- Register (#404)
Merged pull requests:
- Remove unnecessary doc bits after unification (#390) (@ArnoStrouwen)
- Rewrite MOE algo (#394) (@vikram-s-narayan)
- [skip ci] badges (#397) (@ArnoStrouwen)
- doc cov (#398) (@ArnoStrouwen)
- Add Invalidations.yml [skip ci] [skip tests] (#401) (@devmotion)
- add compat for Distributions.jl (#405) (@vikram-s-narayan)
- Update SurrogatesMOE docs (#407) (@vikram-s-narayan)
- only fill upper part of radial basis matrix as it is symmetric (#409) (@kylebeggs)
v6.4.0
Surrogates v6.4.0
Closed issues:
- MOE tests are commented out (#259)
- Setting "strict mode" in docs build throws error (#363)
- Make GEKPLS differentiable (#371)
- Add docstrings for GEKPLS functions (#373)
- [question] Reason for splitting into sub-packages (#384)
Merged pull requests:
- Make GEKPLS differentiable and surrogate optimization compatible (#381) (@vikram-s-narayan)
- Update gekpls.md for new inputs and functionality (#382) (@vikram-s-narayan)
- make docs work with strict mode (#383) (@vikram-s-narayan)
- add MOE to separate sub-package (#386) (@vikram-s-narayan)
- Speed up radial basis functions when output is single number (#392) (@fjebaker)
v6.3.0
Surrogates v6.3.0
Closed issues:
- Performance of kriging compared to mogp-emulator? (#251)
- Radial basis doesn't error with incorrect inputs. (#337)
- Add Hyperparam Optimization for GEKPLS Theta (#366)
- "GEKPLS not defined" error (#367)
- Making Kriging differentiable and better initial hyperparams (#368)
- Kriging Produces Nearly Identical Results for All Inputs After 3 Dimensions (#369)
- How to evaluate surrogate model? (#377)
- Degree modification of multivariate polynomial chaos surrogate model (#378)
Merged pull requests:
- Make theta a user-supplied array param (#370) (@vikram-s-narayan)
- make kriging differentiable (#372) (@archermarx)
- New default hyperparameters and miscellaneous bug fixes for Kriging. (#374) (@archermarx)
- Input dimension checking when evaluating surrogates (#376) (@archermarx)
v6.2.0
Surrogates v6.2.0
Closed issues:
- Exponentially bad performance of radial basis polynomial basis evalution (#264)
- [BUG] Radial basis function interpolation with anything other than linearRadial basis functions is wrong (#316)
- Delete sampling in Surrogates.jl in favor of QuasiMonteCarlo.jl (#338)
- Unsatisfiable requirements detected for package QuasiMonteCarlo (#356)
- What's the purpose of neural network example (#365)
Merged pull requests:
- Add multiquadratic test (#353) (@ranjanan)
- Fix typos in the definition of the sphere function (#358) (@liuyxpp)
- GEKPLS With Custom PLS Based on SKLearn PLS (#359) (@vikram-s-narayan)
- Fix a few typos in the docs (#360) (@pitmonticone)
- setup SciML Style (#361) (@ChrisRackauckas)
v6.1.1
Surrogates v6.1.1
v6.1.0
Surrogates v6.1.0
Closed issues:
- LoadError: UndefVarError: NeuralSurrogate not defined (#344)
- GaussOrthoPoly not defined (#345)
- Parameter optimization (#351)
Merged pull requests:
- Make Expected Improvement find minimum of function (#303) (@archermarx)
- Bump compats of all libs (#342) (@ranjanan)
- Some docstring fixes (#343) (@archermarx)
- update docs for multi-package form (#346) (@vikram-s-narayan)
- Minor fix for "GaussOrthoPoly not defined" error (#347) (@vikram-s-narayan)
- Replace samplers with those from QuasiMonteCarlo.jl (#349) (@archermarx)
- CompatHelper: add new compat entry for QuasiMonteCarlo at version 0.2, (keep existing compat) (#350) (@github-actions[bot])
- Remove samplers from Surrogates.jl in favor of QuasiMonteCarlo.jl (#352) (@archermarx)
- Separate out pages (#354) (@ChrisRackauckas)
v6.0.0
v5.0.1
Surrogates v5.0.1
Closed issues:
Merged pull requests:
- add note about need for hyperoptimization in abstractgps surrogate (#336) (@vikram-s-narayan)
- Remove Flux and PolyChaos as necessary dependencies (#340) (@sharanry)