Generative Adversarial Network-Based Architecture



Welcome to the last article of Deep Learning series in the year 2022.

The latest advancement in deep learning techniques is generative adversarial network (GAN) which consists of a generator and a discriminator. I remember playing car racing games in my childhood. One such game I remember is Need For Speed. We used to have cheat codes for this game, like how to blow the car in front of us, and how to avoid police from catching our player.

This GAN model is exactly like a cheat code in deep learning's language.

The generator is trained using new examples. With large training datasets, it makes the machine believe that the images generated by it are real. The discriminator, on the other hand, decides if the image is fake or real. We can rephrase the sentence as, generator component works on regression-based model, while discriminator component works on classification-based model.


With the coming year 2023, we have a surprise for our readers. Info4EEE team has decided to revive two old article series namely, Biomedical Instrumentation, and Digital Image Processing. These two series are directly or indirectly going to affect our understanding rhythm of deep learning.

See you next year readers. Wish you all a very Happy New Year 2023 🎉😊.


Akshay Juneja authored 15+ articles for INFO4EEE Website on Deep Learning.

No comments

Powered by Blogger.