Learning to Rank Caption Chains for Video-Text Alignment
This addresses the challenge of improving vision-language model alignment for video-text tasks, representing an incremental advancement over existing DPO methods.
The paper tackles the problem of aligning video and text by proposing ranking optimization as an alternative to binary direct preference optimization (DPO), showing it outperforms DPO for long-form content generation and assessment, with results indicating the need for finetuning the vision encoder.
Direct preference optimization (DPO) is an effective technique to train language models to generate preferred over dispreferred responses. However, this binary "winner-takes-all" approach is suboptimal for vision-language models whose response quality is highly dependent on visual content. In particular, a response may still be faithful to the visual inputs even if it is less preferable than an alternative. The standard Bradley-Terry DPO formulation lacks this nuance, upweighting winning responses without sufficient regard for whether the "losing" response still maintains high visual fidelity. In this work, we investigate ranking optimization as an alternative that more precisely situates responses' faithfulness to visual inputs. We focus on video-text alignment using detailed video captions, proposing a method to generate challenging, totally ordered caption chains at scale through repeated caption degradation. Our results show ranking optimization outperforms binary DPO for long-form content generation and assessment, and importantly, we find that these approaches require finetuning of the vision encoder to be effective, challenging the view of DPO as purely a language-reweighting process.