NO

Numerical optimization

---

Monday-Wensday 14-16 12-14 Exercise: Goals: ideas and cot--- Monday-Wensday 14-16 12-14 Exercise:

Goals: - ideas and cotext of MLext of ML - optimization problem - duality Lagrange function, convex optimization - primal-dual methods Nummerical optimization - constraint optimization

Peactical exercices: prgramming.

PreReq: Banach-fixpoint, couchy sequences.

---

Lecture-1

Test_data Sets Linear Regeresoin; least squre problem. finally we have a linear system of equation dreived from an optimization problem. SVM : support Vector Machine Minimizing Function + some extera properties NN : try to mimic brain See an animation of how Neuron works. The stuvture of network is fixes What are the difference between a gpu work on ML and HPC.

---

• Def1: given a finite or sometimes sets of data pairs (x_i,y_i) x_i \in R^m we look for a function: f: R^m --> R^n can be subject to noise y_i=f(x_i) + \psi with expectation that E(\psi)=0 the goal find f_thta : R^m --> R^n s.t data is explained best.

ask for the for defenition of ML

---