Online Iterative Reinforcement Learning from Human Feedback with General Preference Model

• URL: http://arxiv.org/abs/2402.07314v2
• Date: Thu, 25 Apr 2024 04:05:06 GMT
• Title: Online Iterative Reinforcement Learning from Human Feedback with General Preference Model
• Authors: Chenlu Ye, Wei Xiong, Yuheng Zhang, Nan Jiang, Tong Zhang,
• Abstract summary: We study Reinforcement Learning from Human Feedback (RLHF) under a general preference oracle. We consider a standard mathematical formulation, the reverse-KL regularized minimax game between two LLMs for RLHF under general preference oracle. We show that this framework is strictly more general than the reward-based one, and propose sample-efficient algorithms for both the offline learning from a pre-collected preference dataset and online learning.
• Score: 17.900999251247256
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We study Reinforcement Learning from Human Feedback (RLHF) under a general preference oracle. In particular, we do not assume that there exists a reward function and the preference signal is drawn from the Bradley-Terry model as most of the prior works do. We consider a standard mathematical formulation, the reverse-KL regularized minimax game between two LLMs for RLHF under general preference oracle. The learning objective of this formulation is to find a policy so that it is consistently preferred by the KL-regularized preference oracle over any competing LLMs. We show that this framework is strictly more general than the reward-based one, and propose sample-efficient algorithms for both the offline learning from a pre-collected preference dataset and online learning where we can query the preference oracle along the way of training. Empirical studies verify the effectiveness of the proposed framework.

Related papers

• Optimal Design for Reward Modeling in RLHF [83.3614658277817]
  We formalize the reward training model in Reinforcement Learning from Human Feedback. We frame the selection of an effective dataset as a simple regret minimization task. We derive bounds on the simple regret under appropriate assumptions.
  arXiv  Detail & Related papers  (2024-10-22T14:36:44Z)
• Lifelong Personalized Low-Rank Adaptation of Large Language Models for Recommendation [50.837277466987345]
  We focus on the field of large language models (LLMs) for recommendation. We propose RecLoRA, which incorporates a Personalized LoRA module that maintains independent LoRAs for different users. We also design a Few2Many Learning Strategy, using a conventional recommendation model as a lens to magnify small training spaces to full spaces.
  arXiv  Detail & Related papers  (2024-08-07T04:20:28Z)
• Iterative Nash Policy Optimization: Aligning LLMs with General Preferences via No-Regret Learning [55.65738319966385]
  We propose a novel online algorithm, iterative Nash policy optimization (INPO) Unlike previous methods, INPO bypasses the need for estimating the expected win rate for individual responses. With an LLaMA-3-8B-based SFT model, INPO achieves a 42.6% length-controlled win rate on AlpacaEval 2.0 and a 37.8% win rate on Arena-Hard.
  arXiv  Detail & Related papers  (2024-06-30T08:00:34Z)
• On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization [33.331389392270665]
  preference matching (PM) RLHF is a novel approach that aligns large language models with the preference distribution of the reward model under the Bradley--Terry--Luce/Plackett--Luce model. Central to our approach is a PM regularizer that takes the form of the negative logarithm of the LLM's policy probability distribution over responses. For practical implementation, we introduce a conditional variant of PM RLHF that is tailored to natural language generation.
  arXiv  Detail & Related papers  (2024-05-26T07:00:05Z)
• Multi-turn Reinforcement Learning from Preference Human Feedback [41.327438095745315]
  Reinforcement Learning from Human Feedback (RLHF) has become the standard approach for aligning Large Language Models with human preferences. Existing methods work by emulating the preferences at the single decision (turn) level. We develop novel methods for Reinforcement Learning from preference feedback between two full multi-turn conversations.
  arXiv  Detail & Related papers  (2024-05-23T14:53:54Z)
• Fine-Tuning Language Models with Reward Learning on Policy [68.70065254564642]
  Reinforcement learning from human feedback (RLHF) has emerged as an effective approach to aligning large language models (LLMs) to human preferences. Despite its popularity, (fixed) reward models may suffer from inaccurate off-distribution. We propose reward learning on policy (RLP), an unsupervised framework that refines a reward model using policy samples to keep it on-distribution.
  arXiv  Detail & Related papers  (2024-03-28T10:02:10Z)
• Nash Learning from Human Feedback [86.09617990412941]
  We introduce an alternative pipeline for the fine-tuning of large language models using pairwise human feedback. We term this approach Nash learning from human feedback (NLHF) We present a novel algorithmic solution, Nash-MD, founded on the principles of mirror descent.
  arXiv  Detail & Related papers  (2023-12-01T19:26:23Z)
• On the Statistical Benefits of Curriculum Learning [33.94130046391917]
  We study the benefits of Curriculum learning (CL) in the multitask linear regression problem under both structured and unstructured settings. Our results reveal that adaptive learning can be fundamentally harder than the oracle learning in the unstructured setting.
  arXiv  Detail & Related papers  (2021-11-13T14:51:07Z)
• Beyond UCB: Optimal and Efficient Contextual Bandits with Regression Oracles [112.89548995091182]
  We provide the first universal and optimal reduction from contextual bandits to online regression. Our algorithm requires no distributional assumptions beyond realizability, and works even when contexts are chosen adversarially.
  arXiv  Detail & Related papers  (2020-02-12T11:33:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.