$1.$Discuss why heterogeneous mixtures of experts are expected to perform better than homogeneous mixtures

of experts. $(4)$

$2.$ Why is weighted voting used in AdaBoost, instead of allowing each expert in the boosting pipeline to have

the same weight? $(4)$

$3.$ Are boosting algorithms robust to outliers? Motivate your answer. $(4)$

$4.$ Consider an ensemble of k$-$NN algorithms. Why does it make sense to include k$-$NN experts with dierent

values for k$?(5)$

$5.$ How do random forests ensure that the dierent decision trees in the forest exhibit dierent behavior? $(4)$

$6.$ Are regression random forests sensitive to outliers? Motivate your answer. $(4)$

$7.$ Explain how bagging ensemble learning models address the bias$-$variance dilemma.