Multiplying the input vector by the weights matrix is essentially computing a linear combination of them.