#forward-pass

A small network of sigmoid units without thresholds: unit A has input X with weight $W_{XA} = 1$, unit B has input Y with weight $W_{YB} = -1$, and the output unit C has inputs A and B with weights $W_{AC} = 1$, $W_{BC} = 0$. For $X = 60$, $Y = 70$ and target $d = 1$, the output error term is $\delta_C = o_C(1 - o_C)(d - o_C)$ and the update is $\Delta W_{AC} = \delta_C\, o_A$ (learning rate 1). What is $\Delta W_{AC}$? Round to 3 decimals.