An inference algorithm for hidden Markov models which computes the posterior marginals of all hidden state variables given a sequence of observations/emissions. It uses dynamic programming to compute the values in two passes, the first going forward in time and the second backward.

(mnemonic) Used to advise how clocks are altered for daylight saving time: local time goes forward in the spring, and back in fall (autumn).

(idiomatic) To show oneself in the best or most positive way possible; to make a favorable impression.

(idiomatic) A situation in which progress is more than offset by adverse developments.