Mingyuan Zhou

Relative Preference Optimization improves alignment of language models with user expectations.

Exploring how machines create images from text prompts and align with human preferences.

Learn how SiD speeds up image creation while maintaining quality.

A new method enhances language models' efficiency without sacrificing quality.

Introducing an efficient dual-policy system for offline reinforcement learning.

A new method improves language models by learning from real-time data.

A new method enhances Text-to-Image models to align with human preferences.

A new approach combining generative modeling and boosting for better learning outcomes.

New methods improve the speed and quality of text-to-image generation.

A new model enhances graph generation using advanced diffusion techniques.

A look at Score Forgetting Distillation and its impact on generative AI.

Discover how Measurement Optimization transforms image processing for clearer results.

A new approach enhances structure safety with less labeled data.