让机器验证事实时既能准确判断，又能清晰展示推理过程。
DecomposeRL: Learning to Ask Useful, Informative, and Diverse Questions for Semi-Supervised, Traceable Claim Verification
- 用强化学习设计问题生成策略，自动分解复杂声明
- 仅用5000个精选数据就达到86.3%的准确率，超越大模型
- 支持少量标注数据下的半监督训练，适合可信AI研究者
事实验证常面临两难：端到端模型准确但不可解释，分解式方法可追溯却性能不足。本文提出DecomposeRL，通过强化学习框架与多维度奖励机制，实现高精度且可追溯的验证。采用数据精炼流程，将11.5万条未标注事实验证数据压缩为5000条高信息密度样本。在包含生物医学、政治、科学等领域的11个基准测试中，使用70亿参数的DecomposeRL在全监督下达到86.3%域内和69.8%域外平衡准确率，比320亿参数基线和GPT-4.1-mini表现更优；在仅10%标注数据的半监督场景下进一步超越基线。代码、数据与模型已公开。
原文摘要 · Abstract (English)
Claim verification splits between end-to-end classifiers that are accurate but yields no inspectable traces, and decomposition-based methods produce inspectable traces but lag performance on benchmark datasets. We propose DecomposeRL an accurate claim-verifier that produce inspectable traces. DecomposeRL frames decomposition as an RL policy trained with GRPO and a multi-faceted reward ensemble, enabling both fully supervised and semi-supervised learning from unlabeled claims. DecomposeRL addresses the prohibitive training cost of GRPO with a data-curation funnel that distills 115K fact-verification claims into a compact, learning-signal-dense subset of 5K claims. We show that a DecomposeRL-7B policy trained with full supervision on only ~5K curated claims achieves 86.3 in-domain and 69.8 out-of-domain balanced accuracy across 11 claim-verification benchmarks containing biomedical, political, scientific, and general-domain claims. Despite being 4x smaller, it matches 32B baselines and GPT-4.1-mini, and it further outperforms baselines in a semi-supervised setting with only 10% labeled claims data. Code, data, and models are available at https://dipta007.github.io/DecomposeRL
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。