AI RESEARCH
ACPO: Counteracting Likelihood Displacement in Vision-Language Alignment with Asymmetric Constraints
arXiv CS.CV
•
ArXi:2603.22165v1 Announce Type: new While Direct Preference Optimization (DPO) has become the de facto approach for aligning Large Vision-Language Models (LVLMs), it suffers from Likelihood Displacement, where the probability of both chosen and rejected responses collapses. This optimization flaw is especially detrimental in multimodal settings: the erosion of chosen likelihoods -- a failure we term Visual Anchor Collapse -- causes models to abandon visual evidence for strong language priors, precipitating significant hallucinations.