Near the right bottom part, is the notation yhat(Tx-1) correct? Should it be yhat(Ty-1)?
x<t>refers to input to the model at timet.y<t>refers to model output based on input at timet.
Since both of x and y share the same time index t, either notation looks okay to me. They convey the same point that after the 1st model input i.e x<1>, output from previous timestep becomes the model input at the current timestep.
1 Like