Parallel inference of LGSSM in the EM algorithm (+ some bug fixes) - #336

Open

kstoneriv3 wants to merge 20 commits into

probml:mainfrom

kstoneriv3:lgssm_parallel_em

Commits on Aug 11, 2023

calculate cross covariance in parallel inference
kstoneriv3
committed
fcd0101

Commits on Aug 12, 2023

add support for parallel inference and sampling
kstoneriv3
committed
57eb463
fix emission logic of parallel filter
kstoneriv3
committed
792fc2c

Commits on Aug 13, 2023

saving while debugging filtering and smoothing
kstoneriv3
committed
0874e03
fix sampling logic
kstoneriv3
committed
44e13e8
maybe fixed the parallel smoother with dynamic emission with inputs
kstoneriv3
committed
3f89021
get all tests to work
kstoneriv3
committed
221d147
Delete lgssm_learning.py
kstoneriv3
authored
17724cf
undo modification of a notebook for PR
kstoneriv3
committed
b16c1d9
apply black
kstoneriv3
committed
c64ddd6
use same indexing everywhere for LGSSM and apply black to these files
kstoneriv3
committed
73a8270
merge main by adopting the indexing convention of Murphy (2023) for the whole LGSSM codebase
kstoneriv3
committed
24a64cf

Commits on Aug 14, 2023

hopefully the final fix of indices
kstoneriv3
committed
7659d54
update docstring a bit
kstoneriv3
committed
c23b076
remove a duplicated function
kstoneriv3
committed
3b25adc
solve merge conflict of parallel_inference_test.py
kstoneriv3
committed
a818f09
Merge branch 'main' into lgssm_parallel_em
kstoneriv3
committed
2d3a786
black
kstoneriv3
committed
1b19bc7

Commits on Aug 15, 2023

add parallel inference flag to conjugate LGSSM
kstoneriv3
committed
70ab368

Commits on Aug 25, 2023

Merge branch 'main' into lgssm_parallel_em
kstoneriv3
committed
c751635