[Merce V]

Then the fact that the “glare” sense of “lower” has been chosen instead of the “make-low” sense is an error arising from pushing the arctic sentences through the front-end. Should I briefly talk about this in the report or are we more interested in other kinds of errors?

[Merce V]

Thanks, it is much more clear now.

[Merce V]

I also have some trouble understanding the difference between summing over all possible values of Q (Forward algorithm) and choosing the most likely state sequence (Viterbi algoritm). I read your reply but I’m still confused as to why we wouldn’t always want the most likely state sequence?

If we think of Q as a probability distribution, wouldn’t the sum over all possible values of Q (“integrating away”) be equal to 1?

I would also like to have a definition of “stochastic model”. Does this refer to the use of probabilities (Bayesian point of view)?

[Author]

Posts