Emerging Trends in AI - PDF

Emerging Trends Advance AI 1. generative adversarial network (GAN) 2. diffusion based models  What is a generative adversarial network (GAN)? 1. Two neural networks compete in term of accuracy 2. Use a cooperative zero-sum game framework to learn, where one person's gain equals another p...

Emerging Trends Advance AI 1. generative adversarial network (GAN) 2. diffusion based models  What is a generative adversarial network (GAN)? 1. Two neural networks compete in term of accuracy 2. Use a cooperative zero-sum game framework to learn, where one person's gain equals another person's loss. 3. The two neural networks are called the generator and the discriminator. 4. Generator is a convolutional NN and the discriminator is a deconvolutional NN. 5. The goal of the generator is to artificially manufacture outputs that could easily be mistaken for real data. 6. The goal of the discriminator is to identify which of the outputs it receives have been artificially created. 7. Generative models create their own training data. 8. While the generator is trained to produce false data, the discriminator network is taught to distinguish between the generator's manufactured data and true examples. 9. If the discriminator rapidly recognizes the fake data that the generator produces -- such as an image that isn't a human face - - the generator suffers a penalty. 10. As the feedback loop between the adversarial networks continues, the generator will begin to produce higher-quality and more believable output and the discriminator will become better at flagging data that has been artificially created. For instance, a generative adversarial network can be trained to create realistic-looking images of human faces that don't belong to any real person. Structure of a GAN Types of GANs GANs come in a variety of forms and can be used for various tasks. The following are the most common GAN types: 1. Vanilla GAN. ◦ simplest ◦ algorithm tries to optimize the mathematical equation using stochastic gradient descent, which is a method of learning an entire data set by going through one example at a time. ◦ consists of a generator and a discriminator. The classification and creation of generated images is done using the generators and discriminators. 2. Conditional GAN ◦ apply class labels ◦ enables the conditioning of the network with new and specific information ◦ the network receives the images with their actual labels, such as "rose," "sunflower" or "tulip" to help it learn how to distinguish between them. 3. Deep convolutional GAN ◦ DCNN for producing high-resolution image generation for differentiated. ◦ Convolutions are a technique for drawing out important information from the generated data in nabling the network to quickly absorb the essential details. 4. CycleGAN ◦ Learn how to transform between images of various styles. For instance, a network can be taught how to alter an image from winter to summer or from an image of a horse to a zebra. One of the most well-known applications of CycleGAN is FaceApp, which alters human faces into various age groups. 5. StyleGAN ◦ produce photorealistic, high-quality photos of faces, but users can modify the model to alter the appearance of the images that are produced. 6. Super resolution GAN ◦ For a low-resolution image to be changed into a more detailed one. Super-resolution GANs increase the image resolution by filling in blurry spots. Popular use cases for GANs Common use cases of GANs include the following: 1. Filling in images from an outline. 2. Generating a realistic image from text. 3. Producing photorealistic depictions of product prototypes. 4. Converting black and white imagery into colour. 5. Photo translations from image sketches or semantic images that are especially useful in the healthcare industry for diagnoses. In video production, GANs can be used to perform the following: 1. Model patterns of human behaviour and movement within a frame. 2. Predict subsequent video frames. 3. Create a deepfake. Diffusion Based Models 1. Diffusion Models are generative models, meaning that they are used to generate data similar to the data on which they are trained. 2. Fundamentally, Diffusion Models work by destroying training data through the successive addition of Gaussian noise, and then learning to recover the data by reversing this noising process. 3. After training, we can use the Diffusion Model to generate data by simply passing randomly sampled noise through the learned denoising process. Benefits of Using Diffusion Models Image Quality and Coherence Stable Training Privacy-Preserving Data Generation Handling Missing Data Applications of Diffusion Models Text to Image Image to Image Image Search Reverse Image Search Example of a text-to-image diffusion model