Question 3. Given the following training samples from two classes: Class 1: $(1,1)^{t},(2,2^{t},2,0)^{t}$ Class 2: $(0,0)^{t},(1,0)^{t},(0,1)^{t}$ (1) Are the two classes linearly separable? (2) Use the Pseudoinverse Approach to find two linear decision boundaries by using the following two vectors for $\mathbf {b}$. $(1,1,1,1,1,1)^{t}, \quad(1,1,1,1,1,2)^{t}$, (You may use any tool to computer matrix inversion.) (3) Verify if the solutions from (2) can indeed classify the samples correctly. what language is accepted by the npda m = ({90,91,92, {a,b), {a,b,z), q0 z, {92]) with transitions delta(q0, a, z) = {(91, a), (q2, \lambda)] delta(qi,b,a) {(91,b)} \delta(91,b,b) = {(91,b)} \delta(q1, a,b) = {(92, \lambda)] CS. JG. 039