$1$ Combining Algorithm $[5$ points$]$

Ava suggest the following learning algorithm for the expert advice framework with n experts: at each given time t$,$ roll a

fair dice; in case of a $1,$ apply the Majority algorithm at time t $($predict the label output by majority of the experts at

time t$),$ and in case of any other number, apply the Randomized Weighted Majority strategy for time t$.$ Note that, to apply

the Randomized Weighted Majority algorithm, Ava$$s algorithm maintains and updates the weights at all rounds. Specify

whether the regret of Ava$$s algorithm is o$($T$)($i$.$e$.,$ sublinear$)$ or $\backslash$Omega $($T$)($i$.$e$.,$ linear$).$ Justify your answer.