arXiv · cs.AI · Laure Berti-Equille · 4 小时前模型评分46
BP-TD：面向同源 大语言模型（LLM，Large Language Model） 智能体的无监督真值发现方法
Where Is the Truth When AI Agents Disagree? Dependence-Aware Belief Propagation for Unsupervised Truth Discovery (Extended Version)
导读
研究提出 BP-TD，在单币 Dawid-Skene 模型的平均场变分推断中加入按智能体去相关权重，使共享同一基座模型的智能体合计约算一票。在 MMLU 上，13 个本地智能体无标签时准确率达 0.716，高于多数投票的 0.624（160 题）；但去相关仅在最大相关群体弱于其他智能体时有效，且无标签统计量无法区分两种情况。
分类置信度
中 · 待人工复核
- 只有单一媒体来源，需要人工复核。
来源级别：媒体与个人
事件类型：研究发布
证据摘录
- “We propose BP-TD, which adds per-agent decorrelation weights to mean-field variational inference on a one-coin Dawid-Skene model, so that agents sharing a base model count together as about one vote.”
- “With 13 local agents on MMLU, BP-TD beats majority vote without any labels (0.716 vs. 0.624 on 160 questions; 0.707 vs. 0.674 on 10,000) and ranks agents by their true accuracy, which classical estimators invert.”
- “The gain depends on the pool: over 2,852 sub-pools of four benchmarks, decorrelation helps when the largest correlated bloc is weaker than the other agents and hurts when it is the strongest, and we found no label-free statistic that tells the two cases apart.”
- “Labels change the picture quickly: with 20 labelled questions, selecting the single best agent is more accurate than every aggregator we tested on MMLU, ARC and TruthfulQA.”
来源：arXiv · cs.AI · arxiv.org