Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.15595 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918090807181312 |
|---|---|
| author | Xiao, Wenyi Wang, Zechuan Gan, Leilei Zhao, Shuai Li, Zongrui Lei, Ruirui He, Wanggui Tuan, Luu Anh Chen, Long Jiang, Hao Zhao, Zhou Wu, Fei |
| author_facet | Xiao, Wenyi Wang, Zechuan Gan, Leilei Zhao, Shuai Li, Zongrui Lei, Ruirui He, Wanggui Tuan, Luu Anh Chen, Long Jiang, Hao Zhao, Zhou Wu, Fei |
| contents | With the rapid advancement of large language models (LLMs), aligning policy models with human preferences has become increasingly critical. Direct Preference Optimization (DPO) has emerged as a promising approach for alignment, acting as an RL-free alternative to Reinforcement Learning from Human Feedback (RLHF). Despite DPO's various advancements and inherent limitations, an in-depth review of these aspects is currently lacking in the literature. In this work, we present a comprehensive review of the challenges and opportunities in DPO, covering theoretical analyses, variants, relevant preference datasets, and applications. Specifically, we categorize recent studies on DPO based on key research questions to provide a thorough understanding of DPO's current landscape. Additionally, we propose several future research directions to offer insights on model alignment for the research community. An updated collection of relevant papers can be found on https://github.com/Mr-Loevan/DPO-Survey. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_15595 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Comprehensive Survey of Direct Preference Optimization: Datasets, Theories, Variants, and Applications Xiao, Wenyi Wang, Zechuan Gan, Leilei Zhao, Shuai Li, Zongrui Lei, Ruirui He, Wanggui Tuan, Luu Anh Chen, Long Jiang, Hao Zhao, Zhou Wu, Fei Artificial Intelligence Computation and Language Machine Learning With the rapid advancement of large language models (LLMs), aligning policy models with human preferences has become increasingly critical. Direct Preference Optimization (DPO) has emerged as a promising approach for alignment, acting as an RL-free alternative to Reinforcement Learning from Human Feedback (RLHF). Despite DPO's various advancements and inherent limitations, an in-depth review of these aspects is currently lacking in the literature. In this work, we present a comprehensive review of the challenges and opportunities in DPO, covering theoretical analyses, variants, relevant preference datasets, and applications. Specifically, we categorize recent studies on DPO based on key research questions to provide a thorough understanding of DPO's current landscape. Additionally, we propose several future research directions to offer insights on model alignment for the research community. An updated collection of relevant papers can be found on https://github.com/Mr-Loevan/DPO-Survey. |
| title | A Comprehensive Survey of Direct Preference Optimization: Datasets, Theories, Variants, and Applications |
| topic | Artificial Intelligence Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2410.15595 |