Exercise 1 – Fill the Gradient Descent Update (with derivative)
Fill the three blanks to write the general gradient descent step for a parameter vector:
\[
{input3}_{t+1} = {input3}_t - {input1}\, \nabla_{{input3}} \, {input2}({input3}_t).
\]
Scalar version (single parameter \(w\)):
\[ w_{t+1} = w_t - {input1} \, \frac{d}{dw} {input2}(w) \Big\rvert_{w=w_t}. \] Type only the symbols:eta— learning rate,J— objective (loss),theta— parameter vector.