Gpy

Fix

* Gamma prior no assignment after init. [mzwiessele]

* 568, product kernel resolution. [mzwiessele]

* 590. [Max Zwiessele]

Y_normalized was not used for running optimization

* Appveyor comment missing. [mzwiessele]

Other

* Bump version: 1.8.5 → 1.8.6. [mzwiessele]

* Merge pull request 597 from marpulli/devel. [Max Zwiessele]

Allow calculation of full predictive covariance matrices with multipl…

* Allow calculation of full predictive covariance matrices with multiple outputs and normalization. [Mark Pullin]

* Merge pull request 600 from marpulli/plotting_fix. [Max Zwiessele]

Fix visible dimensions for plotting inducing points

* Fix visible dimensions for plotting inducing points. [Mark Pullin]

* Merge pull request 599 from marpulli/grads_efficiency. [Zhenwen Dai]

Make predictive_gradients more efficient

* Make predictive_gradients more efficient. [Mark Pullin]

* Merge pull request 587 from esiivola/feature-multioutput. [Zhenwen Dai]

Merge the implementation of Multioutput kernel

* Changed two function names so that they follow the python naming convention. [Siivola Eero]

* Merge remote-tracking branch 'origin' into feature-multioutput. [Eero Siivola]

* Merge pull request 592 from SheffieldML/sparsegp-normalization. [Zhenwen Dai]

fix: 590

* Merge pull request 589 from apaleyes/devel. [Zhenwen Dai]

Implemented utility function to compute covariance between points in GP Model

* Moved posterior_covariance to Posterior class. [Andrei Paleyes]

* Implemented utility function to compute covariance between points in GP Model. [Andrei Paleyes]

* Changed the structure of multioutput kernel so that it doesn't change the API of Kernels + documented the class. [Eero Siivola]

* Merge remote-tracking branch 'origin/devel' into feature-multioutput. [Eero Siivola]

* Merge pull request 585 from YoshikawaMasashi/devel. [Zhenwen Dai]

modify the MLP kernel equation

* Modify the MLP kernel equation. [masashi yoshikawa]

* Added multioutput kern and tests. [Eero Siivola]

* Multioutput kernel + initial test. [Siivola Eero]

* Multioutput kernel + initial test. [Siivola Eero]

* Change dtype for Python 3 in robot_wirelss. [Neil Lawrence]

* Bump the version: 1.8.4 -> 1.8.5. [Zhenwen Dai]

* Update changelog for 1.8.5. [Zhenwen Dai]

* Merge pull request 579 from SheffieldML/multi_out_doc. [Zhenwen Dai]

Improve the documentation for LVMOGP

* Add type into docstring. [Zhenwen Dai]

* Merge branch 'devel' of github.com:SheffieldML/GPy into multi_out_doc. [Zhenwen Dai]

* Remove non-ascii characters. [Zhenwen Dai]

* Improve the documentation for LVMOGP. [Zhenwen Dai]

* Merge pull request 580 from marpulli/devel. [Zhenwen Dai]

Small correction to doc

* Small correction to doc. [Mark Pullin]

* Merge pull request 578 from pgmoren/devel. [Zhenwen Dai]

Fix EP for non-zero mean GP priors (binary classification)

* Fix EP for non-zero mean GP priors. [Moreno]

* Merge pull request 572 from marpulli/devel. [Alan Saul]

Add mean function functionality to dtc inference method

* Add mean function functionality to dtc inference method. [Mark Pullin]

* Merge pull request 573 from pgmoren/devel. [Zhenwen Dai]

Fix DSYR function (See https://github.com/scipy/scipy/issues/8155)

* Fix DSYR function (See https://github.com/scipy/scipy/issues/8155) [Moreno]

* Merge pull request 574 from alansaul/lyapunov_fix. [Alan Saul]

Fixing scipy=1.0.0 incompatibility of lyapunov discovered in PR 573. Coverage issue should be resolved by PR 575.

* Updated sde_kern to work with scipy=1.0.0. [Alan Saul]

* Merge pull request 575 from SheffieldML/matplotlib_testing. [Alan Saul]

Fixing tests for Matplotlib plotting issue

* Removed ImageComparisonFailure 575. [Alan Saul]

ImageComparisonFailure no longer exists which causes issues with travis testing using the most recent matplotlib

* Figured it must be a matplotlib import error 575. [Alan Saul]

New import matplotlib must be missing a package

* Testing Again 575. [Alan Saul]

* Trying to fix tests for Matplotlib plotting issue. [Alan Saul]

* Merge pull request 526 from msbauer/mlp_extended. [Zhenwen Dai]

added extended version of MLP function

* Fix random seed for reproducible results in tests. [msbauer]

* Updated mapping test to pass gradient checks. [msbauer]

* Update mapping_tests.py. [msbauer]

Remove verbosity again after gradient checks passed without problem with verbosity

* Update mapping_tests.py. [msbauer]

Make output of gradient check verbose to diagnose error

* Added extended version of MLP function with multiple hidden layers and different activation functions. [Bauer]

* Merge pull request 562 from SheffieldML/external-mo. [Zhenwen Dai]

Release the implementation of LVMOGP

* Try to fix the issue with model_tests. [Zhenwen Dai]

* Merge with new changes from devel. [Zhenwen Dai]

* Merge pull request 561 from SheffieldML/deploy. [Max Zwiessele]

Deploy

* Merge pull request 560 from SheffieldML/devel. [Max Zwiessele]

appveyor twine upload error fix

* Merge branch 'deploy' into devel. [Max Zwiessele]

* Merge pull request 558 from SheffieldML/devel. [Max Zwiessele]

Uniform prior fix for other domains

* Merge pull request 559 from SheffieldML/PS-upload-error. [Max Zwiessele]

Update appveyor.yml

* The implementation of SVI-MOGP. [Zhenwen Dai]

Other

* Bump version: 1.8.3 → 1.8.4. [mzwiessele]

* Update appveyor.yml. [Max Zwiessele]

* Merge branch 'devel' of github.com:SheffieldML/GPy into devel. [mzwiessele]

* Merge branch 'deploy' into devel. [Max Zwiessele]

* Merge pull request 557 from SheffieldML/devel. [Max Zwiessele]

Paramz 0.8 update

* Merge pull request 544 from SheffieldML/devel. [Zhenwen Dai]

Release GPy 1.8.x

Fix

* Uniform prior instantiation. [mzwiessele]

* Use paramz 0.8.5. [mzwiessele]

Other

* Bump version: 1.8.2 → 1.8.3. [mzwiessele]

Fix

* Uniform prior tests. [mzwiessele]

* Uniform prior can be positive and negative, depending on lower and upper bound. [mzwiessele]

Other

* Bump version: 1.8.1 → 1.8.2. [mzwiessele]

Fix

* Paramz 0.8. [mzwiessele]

* Creating new kernel was missing active dims. [mzwiessele]

* Naming of ard contribution in ARD plot. [Max Zwiessele]

* Replacing np.power with np.square for efficiency. [Chris Tomaszewski]

* Slight modification to MLP mapping to reduce potential for numpy overflows and unnecessary computation. [Chris Tomaszewski]

Other

* Bump version: 1.8.0 → 1.8.1. [mzwiessele]

* Paramz 0.8. [Max Zwiessele]

* Paramz 0.8. [Max Zwiessele]

* Merge pull request 549 from christomaszewski/devel. [Zhenwen Dai]

fix: slight modification to MLP mapping to reduce potential for numpy…

* Merge pull request 545 from pgmoren/devel. [Zhenwen Dai]

Thanks Pablo for the changes about Basic framework for serializing GPy models. I think we need the corresponding docstrings as developer documentation.

* Basic framework for serializing GPy models. [Moreno]

* Merge pull request 543 from icdishb/devel. [Zhenwen Dai]

Merged in the changes about Input warping using Kumar warping.

Thanks for the contribution, Huibin!

* Input warping using Kumar warping. [Shen]

* Merge branch 'deploy' into devel. [Max Zwiessele]

* Merge branch 'devel' into deploy. [mzwiessele]

* Merge branch 'devel' into deploy. [mzwiessele]

* Merge branch 'deploy' of github.com:SheffieldML/GPy into deploy. [mzwiessele]

* Merge pull request 497 from SheffieldML/devel. [Max Zwiessele]

New fixes for parallel optimize and minor fixes

Fix

* Updated keywords. [mzwiessele]

* Beiwang will add GMM in full. [mzwiessele]

* Kernel tests and variational tests. [mzwiessele]

* Plotting tests for new matplotlib. [mzwiessele]

* Model tests numpy integer error. [mzwiessele]

* Replot with new matplotlib. [mzwiessele]

* Offline plotting workaround with squeezing arrays. [mzwiessele]

* Fixed numpy 1.12 indexing and shape preservation. [mzwiessele]

Other

* Bump version: 1.7.7 → 1.8.0. [mzwiessele]

* Merge pull request 527 from adhaka/ll-surv. [Zhenwen Dai]

merged in the changes about Ll surv

* Merging with the main devel branch. [Akash Kumar Dhaka]

* Correcting weibull- changing parameterisation from f to exp(f) similar to loglogistic. [Akash Kumar Dhaka]

* Change import statements for calling locally. [Akash Kumar Dhaka]

* Fresh branch for these new likelihood functions along with test code- will work atleast with La[lace approximation ... [Akash Kumar Dhaka]

* Merge pull request 542 from SheffieldML/EP_refactor. [Zhenwen Dai]

Merge in the EP refactoring

* Merge pull request 534 from adhaka/EP_refactor. [Zhenwen Dai]

Ep refactor

* Some more merge conflicts resolved for syncing with the current devel branch of main GPy. [Akash Kumar Dhaka]

* Bug fixes in test cases due to changes in api of ep functions.. [Akash Kumar Dhaka]

* Robustified binomial likelihood. [Siivola Eero]

* Minor bugfix. [Siivola Eero]

* Fixed two small lbugs. [Siivola Eero]

* Putting exact_inference_gradients again and calling in instead from ep_gradients which is the genreic method for calculating expected gradients used in ep. [Akash Kumar Dhaka]

* Commiting code for some changes to api for calculating ep_gradients, also fixing some issues with gaussian hermite quadrature, no we have both avaialable ... [Akash Kumar Dhaka]

* Adding file for gaussian-kronrod integration, test cases and calculating gradients of log marginal wrt theta-likelihood params. [Akash Kumar Dhaka]

* Fixing a typo-bug in the last commit for ep test case. [Akash Kumar Dhaka]

* Merging with the gpy devel branch to be in sync with the latest code and make pull request again .. [Akash Kumar Dhaka]

* Bug fixes in test cases due to changes in api of ep functions.. [Akash Kumar Dhaka]

* Robustified binomial likelihood. [Siivola Eero]

* Minor bugfix. [Siivola Eero]

* Fixed two small lbugs. [Siivola Eero]

* Putting exact_inference_gradients again and calling in instead from ep_gradients which is the genreic method for calculating expected gradients used in ep. [Akash Kumar Dhaka]

* Commiting code for some changes to api for calculating ep_gradients, also fixing some issues with gaussian hermite quadrature, no we have both avaialable ... [Akash Kumar Dhaka]

* Adding file for gaussian-kronrod integration, test cases and calculating gradients of log marginal wrt theta-likelihood params. [Akash Kumar Dhaka]

* Fixing a typo-bug in the last commit for ep test case. [Akash Kumar Dhaka]

* Removing pods dependency and a few print commands. [Akash Kumar Dhaka]

* Bug fix to prevent infinite loop because of incorrect stopping condition. [Akash Kumar Dhaka]

* Adding some test cases for EP.. more work needs to be done, these are some high level test cases .. [Akash Kumar Dhaka]

* Refactor EP and EPDTC. [Moreno]

* Added initial version of the refactored EP. [esiivola]

* Added initial version of the refactored EP. [esiivola]

* Added initial version of the refactored EP. [esiivola]

* Bump version: 1.6.1 → 1.6.2. [mzwiessele]

* Fix in sparse_gp_mpi optimizer. [Alex Feldstein]

* Fix for parallel optimization. [Alex Feldstein]

* Changes in EP/EPDTC to fix numerical issues and increase the flexibility of the inference. [Moreno]

Changes to avoid numerical issues and improve the performance:
- Keep value of the EP parameters between calls
- Enforce positivity of tau_tilde
- Stable computation of the EP moments for the Bernoulli likelihood
- Compute marginal in the GP model without directly inverting tau_tilde

Changes to improve the flexibility:
- Add parameter for maximum number of iterations
- Distinguish between alternated/nested mode
- Distinguish between sequential/parallel updates in EP

* Cython in linalg. [Max Zwiessele]

did set cython to working if linalg_cython was importable.

* Bump version: 1.6.0 → 1.6.1. [mzwiessele]

* Bump version: 1.5.9 → 1.6.0. [mzwiessele]

* Provide two classes for plotly plots to remove global variable. [Alex Feldstein]

* Bump version: 1.5.8 → 1.5.9. [mzwiessele]

* Bump version: 1.5.7 → 1.5.8. [mzwiessele]

* Add offline plotting for plotly. [Alex Feldstein]

* Merge pull request 513 from adhaka/EP_refactor. [Zhenwen Dai]

adding some test cases for EP.. more work needs to be done, these are…

* Removing pods dependency and a few print commands. [Akash Kumar Dhaka]

* Bug fix to prevent infinite loop because of incorrect stopping condition. [Akash Kumar Dhaka]

* Adding some test cases for EP.. more work needs to be done, these are some high level test cases .. [Akash Kumar Dhaka]

* Merge pull request 512 from pgmoren/EP_refactor. [Zhenwen Dai]

Refactor EP/EPDTC

* Refactor EP and EPDTC. [Moreno]

* Merge pull request 508 from esiivola/EP_refactor. [Zhenwen Dai]

Bring the ongoing working about EP refactoring

* Added initial version of the refactored EP. [esiivola]

* Added initial version of the refactored EP. [esiivola]

* Added initial version of the refactored EP. [esiivola]