Ultimate Step by Step Guide to Deep Learning Using Python: Artificial Intelligence and Neural Network Concepts Explained in Simple Terms (Ultimate Step by Step Guide to Machine Learning Book 2) by Anis Daneyal

Author:Anis, Daneyal [Anis, Daneyal]
Language: eng
Format: epub
Published: 2020-07-17T16:00:00+00:00

Figure 7.2.34: CNNs overview

7.3 Recurrent Neural Networks (RNN)

RNNs are typically used in data sequences, when time and order play an important role. This can mean time series, prediction of next word in a sentence or when a trend or a pattern is meaningful only when we consider a standard order of data points.

Until now, we examined cases when data is independent of time, like images. Also, our data samples were generally independent from each other; the classification of one photo in a class using a CNN did not have any impact on the classification of any previous or any following image. This indicates that such types of networks (CNNs, feedforwards) have no memory.

On the other hand, RNNs need memory, in order to have a perception of the sequence of data. Memory means that past decisions affect current and future decisions. To implement the memory functionality, the RNN uses a repetition of a single cell.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.