Probability - Part Il Part Il Diffusion models are a class of probabilistic machine learning models that learn to de-noise noisy variations of data. Given a dataset, the first step to training these models is to artificially construct noisy data by adding Gaussian noise to a data-point (called Xg) through a process called \"forward noising\". This process then generates a sequence of noisy data-points, termed as x1 ,X2,Xg,and so until any arbitrary horizon to get x. In this question, you will be dealing with one such forward noising process for a 2D random variable. Below is an equation representing the actual code for adding noise to a 2D variable xt: X= X1+Vl-a- 1 Here ; is asample from a standard normal distribution with mean | and covariance matrix io : . In other words, ; ~ N(0, 1). In a standard diffusion process, we generally use a different for each timestep t. However, for the sake of simplicity, let's assume a to be a constant parameter set toa = 0.19. Numeric 2 points Probability Q7 [2pts] 0.1 0.15 0.12 A Given xp = ey = ,and, = ,the value of xX. x, = . 0.9 0.14 0.01 B The value of A can be expressed in the form of p4 /da where p.4 and qu are relatively prime whole numbers. The value of B can be expressed in the form of pp /aB where pp and gz are relatively prime whole numbers. Answer with the value of p4 + qa + pe + gp