2 papers
cs.LG2026
BA-DPO: Bias-Adjusted Direct Preference Optimization for Language Model Alignment
Antonio Ferrara, Alberto Rumi, Francesco Bonchi
Preference-based alignment methods such as Direct Preference Optimization (DPO) use pairwise preferences labeled by human annotators to fine-tune language models. However, annotato…
cs.CY2024
Fair Pairs: Fairness-Aware Ranking Recovery from Pairwise Comparisons
Georg Ahnert, Antonio Ferrara, Claudia Wagner
Pairwise comparisons based on human judgements are an effective method for determining rankings of items or individuals. However, as human biases perpetuate from pairwise compariso…