|
Dr. Rahul Mishra
Himanshu Sharma
Keywords:
Generative Artificial Intelligence, Large Language Models, Transformers, GANs, VAEs, Diffusion Models, Multimodal AI, Cybersecurity, Generative Models, Artificial Intelligence.
Abstract:
Generative Artificial Intelligence (GenAI) has emerged as a major area of research in Computer Science, enabling machines to generate human-like text, images, audio, video, software code, and other forms of synthetic content. Recent developments in Transformer-based Large Language Models (LLMs), Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and diffusion models have significantly expanded the capabilities of intelligent systems. However, different generative models exhibit substantial differences in architecture, training requirements, generation quality, computational complexity, controllability, and security vulnerabilities. This paper presents a comparative study of major generative AI models, focusing on their underlying architectures, applications, strengths, limitations, and security challenges. The study compares GANs, VAEs, Transformer-based models, and diffusion models using parameters such as generation quality, training complexity, controllability, computational requirements, scalability, and application suitability. A comparative analysis indicates that Transformer-based models are particularly effective for language and multimodal tasks, GANs remain useful for high-quality synthetic data generation, VAEs provide efficient latent-space representation, while diffusion models demonstrate strong performance in high-fidelity image and multimodal generation. The paper also examines emerging issues including hallucination, prompt injection, data leakage, adversarial attacks, copyright concerns, model misuse, and deepfake generation. Finally, future research directions involving efficient architectures, trustworthy GenAI, multimodal intelligence, privacy-preserving learning, model compression, and responsible AI are discussed.
|
|
International Journal of Recent Research and Review
ISSN: 2277-8322
Vol. XIX, Issue 3
September 2026
|
PDF View
PUBLISHED
September 2026
ISSUE
Vol. XIX, Issue 3
SECTION
Articles
|
