arXiv cs.AI· Ryo Takizawa, Tempei Kabayama, Chiemi Yamazaki, Shun Kitamura, Ryo Matsuoka, Yuta Ando, Satoshi Kodera, Norihiko Takeda·· 14 小时前AI 评分26
用监督式 LLM 在语言空间融合异构超声心动图 AI 输出
Integrating Heterogeneous Echocardiography AI Outputs in Language Space: Learnable Late Fusion with a Supervised LLM
AI 导读
研究者提出一种基于监督式 LLM 的晚期融合模块，直接在语言空间聚合不同视图的预测发现陈述与测量值，学习病例级融合函数，无需将各上游模型输出转换为二分类决策、概率或固定长度特征向量。在 646 名患者 859 次检查的患者级三折交叉验证中，该方法在单一框架内整合异构输出，表现与所评估的最佳参考策略相当，并优于未经任务特定训练的通用 LLM。
正文 · 原文
Abstract:View- and target-specific echocardiography AI models capture useful but complementary information and may not achieve sufficient predictive performance in isolation. A learned ensemble is a natural way to combine these outputs, but heterogeneous models typically require each output to be converted into a binary decision, probability, confidence score, or fixed-length feature vector. We developed a supervised large language model (LLM)-based late-fusion module that directly aggregates view-specific predicted finding statements and predicted measurements in language space and learns a case-level fusion function. In patient-level three-fold cross-validation of 859 examinations from 646 patients, the proposed method integrated heterogeneous outputs within a single framework, performed comparably to the best-performing reference strategy evaluated, and outperformed a general-purpose LLM without task-specific training. These results suggest that a supervised LLM can combine multiple upstream models without model-specific output conversion or feature engineering that requires medical knowledge.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.39464 [cs.CL] |
| (or arXiv:2610.39464v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.39464 arXiv-issued DOI via DataCite |
Submission history
From: Ryo Takizawa [view email]
[v1]
Thu, 29 Oct 2026 06:36:39 UTC (606 KB)
来源：arXiv cs.AI · arxiv.org