(Solved): Vectors \( \mathbf{a}^{(1)}=\left[\begin{array}{llll}1 & -1 & 1 & 1\end{arr ...

???????

Vectors \( \mathbf{a}^{(1)}=\left[\begin{array}{llll}1 & -1 & 1 & 1\end{array}\right]^{t} \) and \( \mathbf{a}^{(2)}=\left[\begin{array}{llll}-1 & 1 & -1 & 1\end{array}\right]^{t} \) represent the bit maps shown in Figure P6.14. They need to be autoassociated for input vectors at \( \mathrm{HD}=1 \) from the prototypes stored. Design the following two autoassociators: ASSOCIATIVE MEMORIES (a) a linear autoassociator in the form of a linear associative memory with added thresholding element (TLU) at the output, if needed, but no feedback (b) a recurrent autoassociative memory with asynchronous updating. Compare the performance of both networks by evaluating responses to eight input vectors at \( \mathrm{HD}=1 \) from \( \mathbf{a}^{(1)} \) and \( \mathbf{a}^{(2)} \). The nodes of the memory designed in part (b) should be updated in ascending order starting at node 1.