The application of artificial intelligence (AI) kept expanding over the years as the research progress rapidly with a more advance technologies and methods. Try asking a person beside you “have you heard of generative AI?”. The most possible answer may be a ‘No’ and then followed by the other person to start searching the term on the internet. However, if asked about an application particularly FaceApp, which is able to change a photo of person into a younger or an older self, the answer will be otherwise. In other words, without even knowing, people have actually used or familiar with the application of generative AI. Generally, generative AI is an artificial intelligence algorithms that is used to create new contents from existing contents such as text, images even video or audio files. It applies techniques like generative adversarial networks (GAN), transformers, and variational autoencoders. This post discusses the use of generative AI through generative adversarial networks (GAN).
Generative adversarial networks refer to generative models that involve the use of neural networks pitting against each other. GAN functions to train its parts namely; generator and discriminator in adversary way. In the training, generator will constantly learn to generate plausible data as a negative training examples for the discriminator while the discriminator differentiates the fake data from the real data. A successful training of generator will results in a generated content that is able to fool the discriminator or more specifically saying that that the generator is capable of building fake content that is accurately similar to real source.
There are many applications of GAN that are achieved today and that include; Image-to-image conversion, text-to-image conversion, photos to emojis, super resolution and also the FaceApp that we have discussed in the introduction, face aging. Generative AI through GAN is in fact beneficial in certain aspects such as improving the resolution of images from low to high and restoration of old images. It is however raises issues in security where counterfeiting criminal become uneasy to detect and solve.
E-SPIN Group in the enterprise ICT solutions supply, consulting, project management, training and maintenance for multinational corporations and government agencies across the region E-SPIN do business. Feel free to contact E-SPIN for your enterprise digital transformation initiative, project requirement and inquiry.
Read more about Generative AI as in the link below:
What is generative artificial intelligence or generative AI?
