Source: Deep Learning on Medium
Sequence Modeling using RNN
The feed forward neural nets work on the principle of predicting output at a given data point independently i.e. no dependency from past input. As there is no feedback from past, the network is not suitable for sequence modeling. Furthermore, the tasks where large data processing is involved requires Deep Neural Networks (DNN) and hence the term Deep Sequence Modeling.
Sequential processing needs different neural network architecture. For example, an audio signal can be split into sequences of sound waves, text can be converted into sequence of characters or words, processing medical signals like EKGs, stock trend prediction and genomic data processing.
Read full Article on https://www.ankitaism.com/docs/ebooks/deep-learning/sequence-modeling-using-rnn/
Originally published at https://www.ankitaism.com on January 4, 2020.