ABSOL：与LLM协同的聚合贝叶斯子采样

中文摘要 AI 辅助

大型语言模型越来越多地被用作结构化数据的自然语言接口，但在需要一致的证据条件化、依赖感知推理和不确定性估计的答案时，它们仍然不可靠。贝叶斯网络提供了显式的概率推理层，但从数据中学习有用的结构在大规模下仍然代价高昂且脆弱。我们引入了ABSOL，一个混合的LLM引导的贝叶斯网络结构学习框架，将LLM用作有界的语义引导。在跨越27到1041个节点的五个离散贝叶斯网络基准上，ABSOL是唯一在每个基准上都产生可行图的评估方法，并且在使用GPT-5.4的每个大于27个节点的基准上取得了最高的Edge F_1。四种LLM增强，为统计骨干贡献了互补的语义证据，与非LLM聚合骨干相比，平均将Edge F_1提高了+0.23。互补的事后细化实验表明，这些收益部分依赖于限制LLM对最终结构的权威。总之，这些结果表明，当在统计基础的推理管道中用作有界引导时，语言衍生的语义知识可以显著改进可扩展的概率结构学习。ABSOL的代码可在以下网址获取：此http URL。

英文摘要

Large language models are increasingly used as natural-language interfaces to structured data, yet they remain unreliable when answers require consistent evidence conditioning, dependency-aware reasoning, and uncertainty estimation. Bayesian networks provide an explicit probabilistic reasoning layer, but learning useful structures from data remains costly and fragile at scale. We introduce ABSOL, a hybrid LLM-guided Bayesian network structure-learning framework that uses LLMs as bounded semantic guides. Across five discrete BN benchmarks spanning 27 to 1041 nodes, ABSOL is the only evaluated method to produce a viable graph on every benchmark, and achieves the highest Edge F_1 on every benchmark larger than 27 nodes with GPT-5.4. The four LLM augmentations, which contribute complementary semantic evidence to the statistical backbone, improve Edge F_1 over the non-LLM aggregation backbone by +0.23 on average. Complementary post-hoc refinement experiments suggest that these gains depend in part on limiting the LLM's authority over the final structure. Together, these results show that language-derived semantic knowledge can substantially improve scalable probabilistic structure learning when used as bounded guidance within a statistically grounded reasoning pipeline. The code for ABSOL is available at github.com/megagonlabs/absol-bn.

发表机构

• Megagon Labs

机构由 AI 辅助整理，请以论文原文为准。