arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.02708cs.LGcs.AIcs.CL
收藏

BinaryPPO:用于二分类任务的高效策略优化

BinaryPPO: Efficient Policy Optimization for Binary Classification

University of Toronto(多伦多大学);Vector Institute(向量研究所);Max Planck Institute for Intelligent Systems, Tübingen, Germany(智能系统马克斯·普朗克研究所)

Punya Syon Pandey, Zhijing Jin

AI总结:

BinaryPPO通过置信度加权奖励机制提升二分类任务的准确率,优于传统监督微调方法。

摘要:

监督微调(SFT)是二分类任务(如毒性检测、事实性验证和因果推理)的标准方法。然而,SFT在现实世界中常常因标签噪声、类别不平衡或稀疏监督而表现不佳。我们介绍了BinaryPPO,一种离线强化学习大语言模型(LLM)框架,将二分类重新表述为奖励最大化问题。我们的方法利用一种PPO的变种,结合一种置信度加权奖励函数,该函数惩罚不确定或错误的预测,使模型能够从静态数据集中学习稳健的决策策略,而无需在线交互。在八个领域特定的基准测试和多个架构不同的模型上,BinaryPPO将准确率提高了40-60个百分点,最高达到99%,显著优于监督基线。我们深入分析了奖励塑造、优势缩放和策略稳定性在实现这一改进中的作用。总体而言,我们证明了基于置信度的奖励设计为二分类任务提供了一种稳健的替代方案。我们的代码可在https://github.com/psyonp/BinaryPPO上获取。

英文摘要:

Supervised fine-tuning (SFT) is the standard approach for binary classification tasks such as toxicity detection, factuality verification, and causal inference. However, SFT often performs poorly in real-world settings with label noise, class imbalance, or sparse supervision. We introduce BinaryPPO, an offline reinforcement learning large language model (LLM) framework that reformulates binary classification as a reward maximization problem. Our method leverages a variant of Proximal Policy Optimization (PPO) with a confidence-weighted reward function that penalizes uncertain or incorrect predictions, enabling the model to learn robust decision policies from static datasets without online interaction. Across eight domain-specific benchmarks and multiple models with differing architectures, BinaryPPO improves accuracy by 40-60 percentage points, reaching up to 99%, substantially outperforming supervised baselines. We provide an in-depth analysis of the role of reward shaping, advantage scaling, and policy stability in enabling this improvement. Overall, we demonstrate that confidence-based reward design provides a robust alternative to SFT for binary classification. Our code is available at https://github.com/psyonp/BinaryPPO.

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。

↑