Consider a $1$x$3$ convolution $($as per Deep Learning terminology$)$ kernel o $=[$o$1,$ o$2,$ o$3]=[-12-1].$ Output is generated by stacking $3$ such operators on the $1-$D input data $($i$.$e$.,$ the operator is applied three times on the input$).$ What is the width of the equivalent operator O that will give the same output, when applied only once on the input data?