Summary
- list to array fix in node ordering (details)
- removed testing statements from SING (details)
- multimodal gp plot (details)
- laplacian eigenvector graphs (details)
- modified truncation step to retain always old indices (details)
- left-over changes (details)
- started working on some examples (details)
- working on euler bernoulli (details)
- working on bod (details)
- started with the a sketch of the algorithm (details)
- working on algorithm (details)
- lazy map implementation for BOD (details)
- started algorithm for lazy map in Algorithms (details)
- working on banana (details)
- working on banana (details)
- soft truncation, i.e. nearly linear map works better than hard truncation (which does not work at all but in few cases) (details)
- playing with orders (details)
- working on the generation of the distribution (details)
- setup for Lorenz63 done (details)
- some fixes for sequential data assimilation (details)
- working on linear map approximation (details)
- figuring out what's wrong... (details)
- added testing of gradients for distributions (details)
- added tests for gradients of distributions and tried on Lorenz63 (details)
- working on forward Euler map (details)
- working on Forward Euler (details)
- fixed Lorenz63 and added ODEs discretization maps to the framework (details)
- started implementation of Lorenz 96 (details)
- working on Lorenz96 (details)
- added the possibility of skipping prediction steps by composing Euler maps (details)
- added rmse output for smoothing (details)
- lorenz 96 distribution ready (details)
- starting to generate distribution for Lorenz96 (details)
- dynamics generated correctly (details)
- working on the generator of maps with prescribed sparsity for L96 (details)
- working on the generator of maps with prescribed sparsity for L96 (details)
- working on loading maps for L96 (details)
- working on postprocessing (details)
- working on results for AGU (details)
- working on results for AGU (details)
- added missing files l96 (details)
- fixed documentation (details)
- gumbel example re-run for documentation (details)
- working on time plotting of l63 (details)
- fixed low-rank approximation for outer grad (lis) to not compute multiple likelihoods (details)
- fixed low-rank approximation for outer grad (lis) to not compute multiple likelihoods (details)
- working on reduced basis extraction (details)
- working on reduced basis extraction (details)
- working on low-rank filter (details)
- working on low-rank for filtering (details)
- working on generation of maps for low-rank approximations (details)
- working on generation of maps for low-rank approximations (details)
- experimenting with low-rank (details)
- working on low-rank l96 (details)
- working on low-rank l96 (details)
- working on L96 M2 (details)
- working on L96 M2 (details)
- working on L96 M2 (details)
- working on gradient of L96 model II (details)
- cython implementation of gradient works (details)
- action of hessian works for L96 model 2 (details)
- working on preconditioning (details)
- working on l96 m2 (details)
- working on l96 m2 (details)
- running l96 m2 (details)
- preparing linear maps for low-rank model (details)
- re-factoring some nomenclature for hidden Markov chains (details)
- re-factoring some nomenclature for hidden Markov chains (details)
- modal l96 model 2 done (details)
- fixed tmap-seq-tm (details)
- working on filtering preconditioning (details)
- preconditioning through filter almost works (details)
- working on both l96 and l63 (details)
- filter preconditioning works!! (details)
- prepared filtering preconditioning for l96 (details)
- preconditioning through filtering works for l96 (details)
- mpi_map now updates automatically the counters of the methods called on an object (details)
- l96 with filter preconditioning works. NEXT: remove 1st step of smoother (first lag-1 smoothing) and substitute with the regular step using the 0-th filtering (0-th step inference problem) (details)
- removed skip-first and made 0th step to be included in the final map. So the first map in the composition does not involve a transition (details)
- working on combination of filt preconditioner and low-rank preconditioner (details)
- low-rank preconditioning works with filter preconditioning (details)
- working on lr l96 (details)
- working on l96 (details)
- working on l96 lr (details)
- going.. (details)
- working on banded maps (details)
- fixed unit tests for test_derivatives functions (details)
- now LinearSpanApproximation uses a semilattice for multiindex and coefficients storage (details)
- passing tests (details)
- semilattice constructors exposed in default transport map constructors (details)
- experimenting low-rank subspace ideas (details)
- fixed init coeffs bug (details)
- fixed semilattice data structure for serialization (details)
- fixed bugs in tests (details)
- reshaped tests (details)
- passes all tests (details)
- working on fv linear candidate map (details)
- semilattices integrated into adaptivity. To be tested! The spirit is there though. (details)
- semilattices integrated into adaptivity algorithm. fixes #92 (details)
- changed format of semilattices for linear span (details)
- changed format of semilattices for linear span (details)
- fixed callback parameter minimize kl. (details)
- changed active_set to SortedSet to preserve iteration order after serialization (for parallel computing purposes) (details)
- added LinearSpanVertex to the exposed classes (details)
- fixed bug in LinearMap (defined twice...) (details)
- working on constructor (details)
- started on poisson (details)
- working on runner for poisson (details)
- embedding the deep lazy map construction into the builder cycle (details)
- transformed tmap-tm in the logic of CLI. to be tested by doing... (details)
- restructuring all the building architecture (details)
- restructuring of building architecture done. todo on sequential algorithm as well. (details)
- working on reshaping of builders (details)
- some fixes after big merge. Going to branch off this in order to debug and work on Lazy maps at the same time (details)
- working on the runner for lazy maps (details)
- fixing here and there (details)
- working on the lazy map script (details)
- working on scripts for lazy maps (details)
- working on scripts for lazy maps (details)
- fixing new interface (details)
- testing tmap-deep-lazy-tm (details)
- testing tmap-deep-lazy-tm (details)
- testing... (details)
- script runs (details)
- working on poisson with lazy maps (details)
- working on poisson with lazy maps (details)
- working on poisson problem (details)
- working on taxonomy structure (details)
- added missing file (details)
- fixes for cse presentation on Lazy maps (details)
- ... (details)
- going on with taxonomy (details)
- working on distributions (details)
- ... (details)
- working on taxonomy for distributions (details)
- working on taxonomy (details)
- advancing with distributions (details)
- working on decomposable distributions... (details)
- working on decomposable taxonomy (details)
- working on Maps (details)
- ... (details)
- ... (details)
- added Functionals module (details)
- done with Functionals (details)
- adde Likelihoods (details)
- fixed edge (details)
- Started with the refactoring of the code... Things breaks from here... (details)
- working on Maps module refactoring (details)
- ... (details)
- ... (details)
- started working on TriangularCompoentwiseTransportMap (details)
- ... (details)
- ... (details)
- ... (details)
- working on ListCompositeTransportMapBase (details)
- done with ListCompositeTransportMap (details)
- done with InverseTransportMap (details)
- ... (details)
- working on AffineMaps (details)
- ... (details)
- ... (details)
- ... (details)
- ... (details)
- starting with debugging of Maps (details)
- starting with debugging of Maps (details)
- ... (details)
- unit testing Maps (details)
- working on unittest Maps (details)
- working on unit tests... (details)
- working on unittests (details)
- passes test_transportmaps (details)
- fixed most of unittests... (details)
- some other fixes (details)
- working on distributions (details)
- distributions done (details)
- adding missing file (details)
- working on Functionals (details)
- testing Functionals (details)
- updated distribution (details)
- some fixed (details)
- fixing unittests (details)
- some fixes (details)
- working on reaction diffusion problem (details)
- working on pdes (details)
- fitz hugh (details)
- still picking a pde... (details)
- left two l2 minimization errors... (details)
- passes all tests (details)
- picking pde... (details)
- added missing file (details)
- working on Burgers test problem (details)
- working on burgers; (details)
- working on Burgers (details)
- working on Burgers (details)
- generating distribution (details)
- working on burgers Laplace approximation (details)
- fix committed for mpi with inverse from samples. The mpi_pool is not attached to objects anymore (so that it will be never serialized), but passed along with parameters (details)
- additional fixes (details)
- working on Burgers (details)
- adding Hessian (details)
- added plotting of generated samples (details)
- working on Burgers (details)
- modified gradboot stopping criterion (details)
- updates... (details)
- fixed semilattice handling in adaptation and many other things.. (details)
- adding TC setting for Poisson equation (details)
- working on Poisson (details)
- ... (details)
- adjoint of log-likelihood cost works... (details)
- working on Poisson generation (details)
- working on transforming Functionals into Maps (details)
- moving Functionals into Maps (details)
- updated taxonomy... needs to reflect the fact that Functionals is a submodule of Maps (details)
- fixed taxonomy (details)
- fixing unit tests (details)
- added deprecation message if importing TransportMaps.Functionals (details)
- working on unittests and fixed some docs (details)
- fixed dependencies (details)
- passing unit tests (details)
- added speed test for Poisson (details)
- working on Poisson (details)
- working on Poisson. Introduced action_grad_x in order to speed up in certain cases (details)
- making Functionals into maps with output in one dimension. This will allow the use of all the machinery available for maps (composition etc) also for functionals... (details)
- working on unit tests... (details)
- added title to taxonomy (details)
- all tests pass (details)
- fixed merge with Functionals->Maps. Working on Poisson. Fixed taxonomy. (details)
- RejectionSampler + multi dim uniform (details)
- working on poisson (details)
- working on dimension reduction... (details)
- working on poisson (details)
- working on Cantilever (details)
- adding missing file (details)
- working on timoshenko (details)
- working on banana (details)
- working on lazy with identity tails (details)
- lazy maps have now an identity tail (details)
- working on timoshenko (details)
- working on the generation of the target distribution (details)
- working on Timoshenko distribution (details)
- added puwr ess routine :) (details)
- timoshenko... (details)
- working on timoshenko (details)
- working on timoshenko (details)
- timoshenko set up done (details)
- timoshenko works O_O' (details)
- postprocessing timoshenko (details)
- working on timoshenko (details)
- hard truncation is back.. (details)
- improved action_grad_x with action_adjoint_grad_x to compute gradients left to right if output < input (details)
- improved action_grad_x with action_adjoint_grad_x to compute gradients left to right if output < input (details)
- implementing missing adjoint functions (details)
- Done with LogCox for Lazy maps (details)
- working on cantilever (details)
- staring with Poisson (details)
- added IdentityEmbeddedParametricTransprotMap and made minimize_kl_divergence general for any parametric transport map (details)
- added dolfin version markers/warnings for solvers (details)
- modified ess reporting (details)
- working on timoshenko (details)
- going on with plots (details)
- timoshenko plots (details)
- poisson plots (details)
- preparing code for submission NIPS (details)
- working on Poisson... (details)
- modified poisson distribution (with diagonal likielihood) to allow for lis type model reduction (details)
- Laplace approximation log output to INFO (details)
- test-source.sh added PATH for Jenkins (details)
- update setup (details)
- updating setup (details)
- updating default map call (details)
- working on poisson (details)
- updating to TMv3 (details)
- updating hess x to TM v3 (details)
- updating numerical examples for SING (details)
- fixing tests (details)
- fixed unit tests (details)
- added pip install cython in jenkins script (details)
- added .pyx to the manifest (details)
- updated jenkins script (details)
- working on jenkins script (details)
- passing all tests and updated jenkins scripts (details)
- fixed construction script (details)
- made CYTHON submodules optional (details)
- made CYTHON submodules optional (details)
- workign on random rotations (details)
- working on LogisticRegression model (details)
- working on parkinson (details)
- done likelikhood (details)
- distribution generation done (details)
- working on results for logit problem (details)
- working on differente settings for parkinson (details)
- working on posterior predictive (details)
- mod to logit likelihood (details)
- working on stan example (details)
- working on stan example (details)
- fixed bug in adjoint_grad_x for composite maps (details)
- working on parkinson dataset (details)
- working on parkinson data set (details)
- fixed log-likelihood for logistic regression models with log(1+\eps) approximation (details)
- fixed bug in hess_x_log_pullback function (details)
- renamed aux for windows benefit... (details)
- auxiliary.... (details)