I want help in this question [Last Part in Attached Image]. The attached image contains my attempt.
Let $\{(x_i,y_i)\}_{i=1}^N$ be given with $x_i\in \mathbb R^n$ and $y_i\in \mathbb R$. Assume $N<n$, and $x_i,\;i=1,2,\ldots, N$, are linearly independent. Consider the ridge regression $$\min_{a \in \mathbb R^n} \sum_{i=i}^N(\langle a,x_i\rangle - y_i)^2+\lambda||a||_2^2,$$ where $\lambda \in \mathbb R$ is a regularization parameter, and we set the bias $b=0$ for simplicity.
Denote $$f(\textbf a)= \sum_{i=i}^N(\langle \textbf a,x_i\rangle - y_i)^2+\lambda||\textbf a||_2^2 $$ Calculate $\nabla f(\textbf a).$
My attempt, and entire question
