arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.14874cs.CL

Inverse-Q*:用于无偏好数据对齐大型语言模型的Token级强化学习

Inverse-Q*: Token Level Reinforcement Learning for Aligning Large Language Models Without Preference Data

School of Computer Science, Fudan University(复旦大学计算机科学学院);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)

Han Xia, Songyang Gao, Qiming Ge, Zhiheng Xi, Qi Zhang, Xuanjing Huang

AI总结:

本文提出Inverse-Q*框架,通过优化Token级强化学习且无需额外奖励或价值模型,直接从模型响应中估计条件最优策略,在低资源环境下实现对齐,其收敛速度与对齐效果超越PPO。

摘要:

人类反馈强化学习(RLHF)已被证明在将大型语言模型与人类意图对齐方面是有效的,但它通常依赖于诸如近端策略优化(PPO)等复杂方法,这些方法需要大量的超参数调优,并在样本效率和稳定性方面带来挑战。在本文中,我们介绍了Inverse-Q*,这是一个创新的框架,它通过优化Token级强化学习而无需额外的奖励或价值模型,超越了传统的强化学习方法。Inverse-Q*利用直接偏好优化技术,但通过直接从模型的响应中估计条件最优策略来扩展这些技术,从而促进了更细粒度和灵活的策略塑造。我们的方法减少了对人类标注和外部监督的依赖,使其特别适合低资源环境。我们提供了广泛的实验结果,证明Inverse-Q*不仅在收敛速度和模型响应与人类偏好对齐方面与PPO相当,而且有可能超越PPO的有效性。我们的研究结果表明,Inverse-Q*为传统的RLHF方法提供了一种实用且稳健的替代方案,为更高效和适应性强的模型训练方法铺平了道路。

英文摘要:

Reinforcement Learning from Human Feedback (RLHF) has proven effective in aligning large language models with human intentions, yet it often relies on complex methodologies like Proximal Policy Optimization (PPO) that require extensive hyper-parameter tuning and present challenges in sample efficiency and stability. In this paper, we introduce Inverse-Q*, an innovative framework that transcends traditional RL methods by optimizing token-level reinforcement learning without the need for additional reward or value models. Inverse-Q* leverages direct preference optimization techniques but extends them by estimating the conditionally optimal policy directly from the model's responses, facilitating more granular and flexible policy shaping. Our approach reduces reliance on human annotation and external supervision, making it especially suitable for low-resource settings. We present extensive experimental results demonstrating that Inverse-Q* not only matches but potentially exceeds the effectiveness of PPO in terms of convergence speed and the alignment of model responses with human preferences. Our findings suggest that Inverse-Q* offers a practical and robust alternative to conventional RLHF approaches, paving the way for more efficient and adaptable model training approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑