Generative Adversarial Networks or GANs are used in drawing or sketching some figures, to be precise it is used for generating an output that is fairly distinctive from the input dataset.
We know that Neural Networks are used for classification, Given an input dataset it predicts the most possible category with the help of the learned features during the training. For instance, given an input dataset as images of dog and cat to the network, somewhere in the complete network it learns the distinctive features of the cats and dogs.
So given a new input image of a cat the neural network gives output as 1 or 0 (1 : cat and 0: dog).
But can the same network generate a distinctive image of a cat?
The answer is No, the neural networks can only classify whether it is a cat or some other animal. GAN’s or Generative Adversarial Network cover it up here.
GAN’s tends to generate a unique output after learning from the input dataset
Hence we can define it as,
an unsupervised learningtask in machine learning that involves automatically discovering and learning the regularities or patternsin input data in such a way that the model can be used to generate or output new examples that plausibly could have been drawn from the original dataset. Generative Adversarial Networks or GANs are a deep-learning-based generative model.
The GANs have two sub-models
1. Discriminator model: classify examples as either real (from the
domain) or fake (generated)
2. Generator model: given a random noise gives an new output from that.
The reward of the Generator is inverse to that of the discriminator. For every output classified as fake by the discriminator, the generator gets a negative reward as they are not trained to do that, as when the generator receives a negative reward it tries improve its performance until it receives a positive reward.
2. How AI Will Power the Next Wave of Healthcare Innovation?
3. Machine Learning by Using Regression Model
4. Top Data Science Platforms in 2021 Other than Kaggle
Hence it is a fact that
“A really good generative model may be able to generate new examples that are not just plausible, but indistinguishable from real examples from the problem domain”
GANs are are based on a game theoretic scenario in which the generator network must compete against an adversary. The generator model directly produces samples. Its adversary, the discriminator model attempts to distinguish between samples drawn from the training data and samples drawn from the generator.
Thank you for reading along, we hope you got inspired and loved it. Let us know in the comments your view on the above topic.
Read our recent blogs on Artificial Intelligence and other topic
📌The Blue Brain Project
📌Women In A.I.
📌Computer Language & memory 101!
📌Artificial Intelligence & its impacts
📌How to implement a data quality process
📌What, why’s and who’s of Cyber Security
Don’t forget to give it a clap, that motivates us.
For more, follow us twitter @theaithing and on instagram @theaithing
If you also want to join our community or want to contribute drop a mail on firstname.lastname@example.org