The current versions of information form message passing for LGSSMs use a hybrid representation of dynamics parameters and information form emission parameters. It is often helpful to have routines that operate directly on the block tridiagonal precision matrix and the precision-weighted mean. Examples of such implementations are:
The current versions of information form message passing for LGSSMs use a hybrid representation of dynamics parameters and information form emission parameters. It is often helpful to have routines that operate directly on the block tridiagonal precision matrix and the precision-weighted mean. Examples of such implementations are:
(I've also written information form samplers and smoothers in JAX, along the lines of the last two links.)
We should add similar inference routines to the LGSSM info inference code.