#viterbi

Question 1: Select all that apply

mediumAI-checkedProbabilistic & Graphical Models

In the forward algorithm for an HMM, $\alpha_i(t)$ is the value stored for state $i$ at time $t$ in the trellis. Which statements are true? Select all that apply.

The forward algorithm assumes that the observation sequence and the model parameters are fixed and known1
$\alpha_i(t)$ is the total probability of seeing the observations up to time $t$ and ending in state $i$ at time $t$ , summed over all paths that lead there2
$\alpha_i(t)$ uses no information about the observations after time $t$ 3
The forward algorithm ends by summing the $\alpha$ values of the last time step4
$\alpha_i(t)$ is the probability of the single most likely path that ends in state $i$ at time $t$ 5

Timed test