Creative Generation

Creative generation research focuses on developing computational methods to produce novel and diverse outputs, such as images and text, often leveraging diffusion models and large language models (LLMs). Current efforts concentrate on improving the creativity and diversity of generated content while mitigating biases and preventing the reproduction of training data, exploring techniques like propulsive energy diffusion and investigating the trade-offs between alignment and creative output in LLMs. This research has significant implications for various fields, including advertising, search engines, and artistic practice, by enabling personalized content creation and enhancing human-AI collaboration.