Suppose you’re a chef and also you need to determine every single day what you prepare dinner primarily based on what you cooked yesterday (in fact, you don’t need to repeat) and the way at this time’s climate is!
You determine to take the assistance of a neural network for on a regular basis planning
Right here’s a easy neural network with an enter layer having 9 inputs, a hidden layer and an output layer.
Discover a few issues on this neural community:
- Right here, we outline the variety of inputs beforehand, and this stays fixed.
- Additionally, all inputs are fed to the community on the identical time to start with.
How can we use this type of structure to determine the meals to prepare dinner primarily based on yesterday’s meal?
For that, we’d like a reminiscence of earlier output, which isn’t current in neural networks. Additionally, we can not present enter on at this time’s climate for every day as inputs are handed beforehand.
It is a easy drawback the place our Recurrent Neural Networks can come to the rescue. An RNN has a reminiscence which takes choices primarily based on earlier outputs and present enter.
Within the beneath diagram, at every time step, an enter is given, an output is produced, and the identical output is…
