ToE: A Hierarchical and Explainable Claim Verification Framework with Dynamic Multi-source Evidence Retrieval and Aggregation

ToE：一种具有动态多源证据检索与聚合功能的层级化可解释声明验证框架

The rapid spread of fake news poses increasing threats to information ecosystems, especially as AI-generated misinformation under Generative Engine Optimization (GEO) poisoning allows adversarially crafted content to be systematically surfaced by retrieval systems, contaminating LLM reasoning. 虚假新闻的迅速传播对信息生态系统构成了日益严重的威胁，特别是在生成式引擎优化（GEO）投毒攻击下，AI 生成的虚假信息使得对抗性内容能够被检索系统系统性地呈现，从而污染大语言模型（LLM）的推理过程。

In this paper, we propose Tree of Evidence (ToE), a hierarchical evidence reasoning framework for automated fact-checking that models each claim as a dynamically expanding argument tree. 在本文中，我们提出了“证据树”（Tree of Evidence, ToE），这是一种用于自动化事实核查的层级化证据推理框架，它将每个声明建模为一棵动态扩展的论证树。

ToE integrates a reinforcement learning-driven multi-source retrieval agent, an evidence evaluation agent, and an argument tree aggregation algorithm to iteratively decompose, retrieve, and verify claims through an explainable evidence chain. ToE 集成了一个由强化学习驱动的多源检索智能体、一个证据评估智能体以及一种论证树聚合算法，通过可解释的证据链对声明进行迭代式的分解、检索和验证。

We further provide a theoretical analysis of the retrieval process, deriving a formal error bound that guarantees the learned policy converges to a neighborhood of the information-theoretically optimal policy. 我们进一步对检索过程进行了理论分析，推导出了一个形式化的误差界限，该界限保证了所学习的策略能够收敛到信息论最优策略的邻域内。

Experiments across multiple datasets and backbone LLMs demonstrate that ToE achieves improvements ranging from 4 to 24 percentage points over competitive baselines, with particularly pronounced gains on adversarially poisoned inputs. 在多个数据集和主流大语言模型上的实验表明，与现有的竞争基准相比，ToE 实现了 4 到 24 个百分点的性能提升，在对抗性投毒输入上的增益尤为显著。