Okay so I just did the quiz for week 1 of course 5 of the Deep learning course, and one of the questions was whether Gamma update and Gamma forget in an LSTM play a similar role to Gamma update and 1 minus Gamma update in a GRU. I said yes because these two values determine how C is determined with respect to C~ and C. However, that was apparently wrong because an LSTM uses Gamma forget instead of 1 minus Gamma update. I think this question is very misleading because Gamma update and 1 minus Gamma update are performing the same role in a GRU that Gamma update and Gamma forget perform in an LSTM.

1 Like

Thanks for your report.

+1

Agreed this needs editing.