preference optimization

Preference optimization is a process used to align AI models' outputs with human preferences or desired outcomes. It often involves collecting user feedback and iterating on model training to better serve the specific needs of users.

16 papers