Skip to main content
. Author manuscript; available in PMC: 2021 Mar 11.
Published in final edited form as: Proc IEEE Int Conf Acoust Speech Signal Process. 2018 Sep 13;2018:6533–6537. doi: 10.1109/icassp.2018.8462637

Fig. 2.

Fig. 2.

Modified network used in training illustrated in Algorithm 1. The main difference between the unrolled network in Fig.1.(b) is that the variables Qn; n = 0, .., N are assumed to be fixed. These variables are estimated in the outer loop as shown in Algorithm 1.