Page 204 - Ai_V1.0_Class9
P. 204

Training      Generative AI models are often trained     Conventional AI models are typically
                              using techniques such as generative        trained using supervised, unsupervised, or
                              adversarial networks (GANs), variational   reinforcement learning techniques.
                              autoencoders (VAEs), or autoregressive
                              models.

                Dataset       Generative AI models typically require large  Conventional AI models rely on smaller, more
                              amounts of diverse and representative      curated datasets that are tailored to the task
                              data to learn effectively. These datasets   at hand.
                              often contain thousands or even millions
                              of examples across various categories or
                              classes.

                Output        Generative AI output is fresh, innovative,   Conventional AI produces more predictable
                              and often unexpected.                      output based on existing data.

                Applications  Generative AI is used in the fields of art,   Conventional AI is used in banking, healthcare,
                              music, literature, gaming, and design.     image recognition, and language processing.


                       Types of Generative AI


              Generative AI comes in a variety of forms, each with unique advantages and uses. Let us learn about some of the
              most typical varieties of Generative AI.

              Generative Adversarial Networks (GANs)

              GANs are neural networks that work to produce fresh data. It is made up of two neural networks, which work
              together in a unique adversarial process to create realistic synthetic data. These two neural networks are as
              follows:
                 • Generator Network: It produces data that is as close as possible to real data.
                 • Discriminator Network: It analyses data and provides feedback, i.e. it takes real data and the data generated
                by the generator as input and attempts to distinguish between the two.
              These two networks work together in a cycle where the generator tries to create realistic fake data, and the
              discriminator tries to identify whether the data is real or fake. This back-and-forth process helps the generator
              improve and produce more convincing data over time.

              Some of the examples of GANs are as follows:
                 • It can create portraits of non-existing people.
                 • It can convert images from day to night.

                 • It can generate images based on textual description, for example, if we give a description of a bird, then it will
                create an image that is similar to the description.

                 • It can generate realistic video, which can be used in film production, video games, and generating synthetic data
                for training other AI models, etc.

              Variational Autoencoders (VAEs)

              A variational autoencoder (VAE) is a Generative AI algorithm that uses deep learning to generate new content,
              detect anomalies, and remove noise. This is another class of generative models.

                    202     Artificial Intelligence Play (Ver 1.0)-IX
   199   200   201   202   203   204   205   206   207   208   209