paper-with-me

Papers

COPR: Continual Human Preference Learning via Optimal Policy Regularization

2024-02-22 · Han Zhang, Lin Gui, Yu Lei, Yuanzhao Zhai, Yehong Zhang, Yulan He, Hui Wang, Yue Yu, Kam-Fai Wong, Bin Liang, Ruifeng Xu

Reinforcement Learning from Human Feedback (RLHF) is commonly utilized to improve the alignment of Large Language Models (LLMs) with human preferences. Given the evolving nature of human preferences, continual alignment becomes more crucial and practical in comparison to traditional static alignment. Nevertheless, making RLHF compatible with Continual Learning (CL) is challenging due to its complex process. Meanwhile, directly learning new human preferences may lead to Catastrophic Forgetting (CF) of historical preferences, resulting in helpless or harmful outputs. To overcome these challenges, we propose the Continual Optimal Policy Regularization (COPR) method, which draws inspiration from the optimal policy theory. COPR utilizes a sampling distribution as a demonstration and regularization constraints for CL. It adopts the Lagrangian Duality (LD) method to dynamically regularize the current policy based on the historically optimal policy, which prevents CF and avoids over-emphasizing unbalanced objectives. We also provide formal proof for the learnability of COPR. The experimental results show that COPR outperforms strong CL baselines on our proposed benchmark, in terms of reward-based, GPT-4 evaluations and human assessment. Furthermore, we validate the robustness of COPR under various CL settings, including different backbones, replay memory sizes, and learning orders.

📄 PDF Abstract BibTeX arXiv:2402.14228

Code (0)

등록된 구현이 없습니다.

Tasks

Continual Learning

Methods 이 논문이 사용한 방법론

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Adam 설명 없음
Attention 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Multi-Head Attention 설명 없음

Similar Papers 제목 키워드 기반

COPR: Continual Learning Human Preference through Optimal Policy Regularization

2023-10-24 · Han Zhang, Lin Gui, Yuanzhao Zhai, Hui Wang 외

The technique of Reinforcement Learning from Human Feedback (RLHF) is a commonly employed method to improve pre-trained Language Models (LM), enhancing their ability to conform to human preferences. Nevertheless, the cur…

Continual Learningreinforcement-learningReinforcement Learning

CPPO: Continual Learning for Reinforcement Learning with Human Feedback

2024-01-16 · Conference 2024 1 · Han Zhang, Yu Lei, Lin Gui, Min Yang 외

The approach of Reinforcement Learning from Human Feedback (RLHF) is widely used for enhancing pre-trained Language Models (LM), enabling them to better align with human preferences. Existing RLHF-based LMs however req…

Continual Learningreinforcement-learningReinforcement Learning

Coprocessor Actor Critic: A Model-Based Reinforcement Learning Approach For Adaptive Brain Stimulation

2024-06-10 · Michelle Pan, Mariah Schrum, Vivek Myers, Erdem Biyik 외

Adaptive brain stimulation can treat neurological conditions such as Parkinson's disease and post-stroke motor deficits by influencing abnormal neural activity. Because of patient heterogeneity, each patient requires a u…

Model-based Reinforcement Learningreinforcement-learningReinforcement Learning

Reinforcement Learning From Imperfect Corrective Actions And Proxy Rewards

2024-10-08 · Zhaohui Jiang, Xuening Feng, Paul Weng, Yifei Zhu 외

In practice, reinforcement learning (RL) agents are often trained with a possibly imperfect proxy reward function, which may lead to a human-agent alignment issue (i.e., the learned policy either converges to non-optimal…

Atari GamesAutonomous Drivingreinforcement-learningReinforcement Learning+1

FAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement

2026-07-01 · Haoran Hao, Shahram Najam Syed, Jeffrey Ichnowski, Jeff Schneider arxiv

Robot policies inevitably encounter failures when deployed in real environments. Naive retries often repeat the same mistakes, while many existing recovery methods rely on human intervention. In this paper, we propose Fa…