bradley-terry model
The Bradley-Terry model is a statistical model used to estimate the probability of one item winning over another based on pairwise comparisons, applicable in ranking and recommendation systems.
- Doubly Robust Alignment for Large Language Models
- Greedy Sampling Is Provably Efficient For RLHF
- Improving LLM General Preference Alignment via Optimistic Online Mirror Descent
- On Extending Direct Preference Optimization to Accommodate Ties
- Preference-Based Dynamic Ranking Structure Recognition
- Risk-aware Direct Preference Optimization under Nested Risk Measure