Summary of Direct Preference Optimization Of Video Large Multimodal Models From Language Model Reward, by Ruohong Zhang et al.
Direct Preference Optimization of Video Large Multimodal Models from Language Model Reward
by Ruohong Zhang, Liangke Gui, Zhiqing Sun, Yihao Feng, Keyang Xu, Yuanhan Zhang, Di Fu, Chunyuan Li, Alexander Hauptmann, Yonatan Bisk, Yiming Yang
First submitted to arxiv on: 1 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research paper proposes a novel framework for enhancing the generalization abilities of large language models in video instruction-following and preference modeling tasks. The approach utilizes detailed video captions as proxy content, allowing language models to incorporate this information as supporting evidence for scoring video Question Answering predictions. This tailored reward mechanism demonstrates robust alignment with OpenAI GPT-4V model’s input framework. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper improves the performance of large language models in video instruction-following and preference modeling tasks by using detailed video captions as proxy content. The approach shows that incorporating this information can significantly improve the accuracy of video Question Answering predictions, making it a valuable contribution to the field. |
Keywords
» Artificial intelligence » Alignment » Generalization » Gpt » Question answering
