Название: Generative AI models
Автор: Jovan Pehcevski
Издательство: Arcler Press
Год: 2024
Страниц: 419
Язык: английский
Формат: pdf (true)
Размер: 50.2 MB

The Generative AI is especially powerful in several areas, such as text generation (product description, article writing), image and video generation (AI-generated pictures and videos for marketing industry), and voice and sound generation (for film industry). This book edition covers different topics of generative AI models, including: image generation techniques, video generation techniques, speech / voice generation techniques, and societal and ethical issues of these models.

The world is changing rapidly, and scientific and technological progress plays a key role in this. Artificial Intelligence (AI), which has permeated all spheres of social, economic, scientific research and everyday life, has a special influence. Today, the spotlight is on generative artificial intelligence, which has the potential to change the world in the coming years, in terms of development, commercial and social perspectives.

Within a few months, ChatGPT has become the most prominent representative of the new generation of generative artificial intelligence systems. Others are called LaMDA, DALL-E or Stable Diffusion. These programs produce fundamentally new texts, codes, images or even videos. The results are so convincing that it is often impossible to tell whether they were created by human or machine.

The generative AI is especially good and applicable in 3 major areas:
• Text generation - applications in law (drafting contracts), medicine (diagnostics), journalism (news production), education (production of educational materials), science (search and generation of scientific papers), etc.
• Image and video generation - application in marketing (advertisement creation), media (virtual host), art, architecture, design, social networks, etc.
• Voice and sound generation - application in the film industry (special effects), music industry, customer support (virtual references), surveying (automation of telephone surveys), etc.

Generative Artificial Intelligence is expected to revolutionize the way people work or find information online. But it also raises numerous ethical questions, as rarely has any technology done so far. There are fears that millions of people could lose their jobs, or that the system could be misused for disinformation, or even that the world, as we know it, will end. This sparked a debate about the necessary rules and regulation of AI.

Section 1 focuses on image generation techniques, describing image generation and style transfer algorithm based on deep learning; an overview of image caption generation methods; an application of an improved DCGAN for image generation; a private face image generation method based on deidentification in low light; and a remote sensing image data scene generation method in smart city.

Section 2 focuses on video generation techniques, describing realistic speech-driven talking video generation with personalized pose; video transformation in big video era and its impact on content editing; a fast depth-map generation algorithm based on motion search from 2D video contents; and adaptive content management for UGC video delivery in mobile internet era.

Section 3 focuses on voice and speech generation, describing generating the voice of the interactive virtual assistant; voice quality modelling for expressive speech synthesis; prosodically rich speech synthesis interface using limited data of celebrity voice; and an overview of resources for development of Hindi speech synthesis system.

Section 4 focuses on societal and ethical issues of generative AI models, describing AI-human symbiotes reinventing innovation and what the new centaurs will mean for cities; the impact of AI and automation on new jobs; a development of artificial intelligence in taxation; AI films as reflections on reality and illusion; and an ecologically sound procedural generation of natural environments.








[related-news] [/related-news]