Content Generation
Content generation research focuses on developing computational methods to automatically create various forms of media, including text, images, videos, and 3D models, often aiming for human-level quality and controllability. Current research emphasizes improving model architectures like large and small language models (LLMs and SLMs), diffusion models, and retrieval-augmented generation (RAG), often combining these approaches for enhanced performance and addressing issues like hallucination and style consistency. This field is significant due to its potential to automate content creation across numerous domains, impacting fields from marketing and entertainment to education and scientific research, while also raising important ethical considerations regarding authorship, bias, and the detection of AI-generated content.
Papers
MARVEL-40M+: Multi-Level Visual Elaboration for High-Fidelity Text-to-3D Content Creation
Sankalp Sinha, Mohammad Sadil Khan, Muhammad Usama, Shino Sam, Didier Stricker, Sk Aziz Ali, Muhammad Zeshan Afzal
Evaluating Generative AI-Enhanced Content: A Conceptual Framework Using Qualitative, Quantitative, and Mixed-Methods Approaches
Saman Sarraf