Summary of Lrhp: Learning Representations For Human Preferences Via Preference Pairs, by Chenglong Wang et al.
LRHP: Learning Representations for Human Preferences via Preference Pairs
by Chenglong Wang, Yang Gan, Yifu Huo, Yongyu Mu, Qiaozhi He, Murun Yang, Tong Xiao, Chunliang Zhang, Tongran Liu, Jingbo Zhu
First submitted to arxiv on: 6 Oct 2024
Categories
- Main: Computation and Language (cs.CL)
- 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 aims to improve training for aligning human preferences with AI models by developing a new framework called LRHP (Learning Representations for Human Preferences) that goes beyond traditional reward modeling. The authors introduce a preference representation learning task and demonstrate its effectiveness in two downstream tasks: selecting relevant data and predicting the margin between preferred and dispreferred options. By leveraging rich and structured representations of human preferences, the model achieves strong performance in these tasks, outperforming baselines. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps AI learn what humans like or dislike better by creating a new way to represent human preferences. Instead of giving a single number as a reward signal, this approach represents preferences more accurately and in a structured way. The authors test their method on two tasks: picking the right data and predicting what people prefer. By using these rich representations, they show that AI can perform better than before. |
Keywords
» Artificial intelligence » Representation learning
