6. In class we showed that Adaboost can be viewed as learning an additive model via functional gradient descent to optimize the following exponential loss function: exp(-9. hi () Our derivation showed that in each iteration I, to minimize this objective, we should seek an hi that minimizes the weighted training error, where the weight of each example w/ exp(-yillah(2:)) prior to normalization. Show how this definition of w is proportional to the Dr(i) computed by Adaboost. 6. In class we showed that Adaboost can be viewed as learning an additive model via functional gradient descent to optimize the following exponential loss function: exp(-9. hi () Our derivation showed that in each iteration I, to minimize this objective, we should seek an hi that minimizes the weighted training error, where the weight of each example w/ exp(-yillah(2:)) prior to normalization. Show how this definition of w is proportional to the Dr(i) computed by Adaboost