发表机构
The Overlake School; Stanford University（奥弗克学校; 斯坦福大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对虚假信息分流，提出基于前30分钟传播树预测后续增长的方法，结合回复模式与证据检索，构建人机协同审核系统以提升识别效率。
AI 中文摘要
有限的审核团队必须在新兴主张的最终传播范围已知之前，识别出哪些主张可能持续增长。我们将早期虚假信息分流聚焦于这一持续性预测问题：根据活动的前30分钟预测后续记录的传播树增长。在FibVID数据集上，我们比较了早期节点数量与结构深度熵、时间到达熵及其组合，同时将每个原始主张的所有传播树保留在同一分区中。在来自59个主张组、与训练集分离的352棵测试树上，组合模型将对数变换后未来增长的$R^2$从0.307提高到0.323，并将对数平均绝对误差（log-MAE）降低了2.4%（95%主张自助法置信区间，-0.4%至5.2%）。这一提升在97棵高活动性传播树中尤为显著：$R^2$从0.248升至0.395，Spearman相关系数$\ ho$从0.394升至0.529，log-MAE下降11.7%（95%置信区间，-1.9%至23.9%）。作为30分钟增长预测的补充，我们分析了563个PHEME线程中的前15条回复。在该队列中，第15条回复的中位到达时间为28.8分钟；52.0%的线程在30分钟内达到固定回复前缀，71.6%在一小时内达到。即使没有LLM生成的事实准确性维度，剩余的立场、交际和情感状态组合仍保留了跨事件排序信号（ROC-AUC 0.538）；加入该维度后ROC-AUC提升至0.562。在另一项针对229条主张的Check-COVID评估中，倒数排名融合（reciprocal-rank fusion）在74.2%的主张中于前五名内检索到黄金证据文档，在94.3%的主张中于前20名内检索到；句子重排序达到Recall@20为58.1%。我们提出一个集成的人机审核系统，将这些早期预测、响应模式和检索到的证据结合起来，用于依赖主张潜在病毒式传播的虚假信息分流。
英文摘要
Limited review teams must identify which emerging claims are likely to keep growing before their eventual reach is known. We center early misinformation triage on this continuation-forecasting problem: predicting subsequent recorded propagation-tree growth from the first 30 minutes of activity. On FibVID, we compare early node count with structural depth entropy, temporal arrival entropy, and their pair while keeping all propagation trees from each original claim in one partition. Across 352 test trees from 59 claim groups separate from training, the combined model raises $R^2$ for log-transformed future growth from 0.307 to 0.323 and reduces log-MAE by 2.4% (95% claim-bootstrap CI, -0.4% to 5.2%). The gain is especially pronounced among 97 high-activity trees: $R^2$ rises from 0.248 to 0.395, Spearman's $ρ$ from 0.394 to 0.529, and log-MAE falls by 11.7% (95% CI, -1.9% to 23.9%). Complementing the 30-minute growth forecast, we analyze the first 15 replies in 563 PHEME threads. In this cohort, the 15th reply arrives after a median of 28.8 minutes; 52.0% reach the fixed reply prefix within 30 minutes and 71.6% within one hour. Even without the LLM-generated factual-accuracy dimension, the remaining stance, communicative, and affective state composition retains cross-event ranking signal (ROC-AUC 0.538); including that dimension increases ROC-AUC to 0.562. In a separate Check-COVID evaluation of 229 claims, reciprocal-rank fusion retrieves a gold evidence document within the top five for 74.2% of claims and within the top 20 for 94.3%; sentence reranking reaches Recall@20 of 58.1%. We propose an integrated human-review system that brings these early forecasts, response patterns, and retrieved evidence together for misinformation triage relying on the potential virality of claims.
CommentsWork done as part of the SHTEM Internship at Stanford University
