English(EN) LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations

LLM框架LIMSSR解决训练数据不完整情况下的多模态学习问题

作者 PulseAugur 编辑部 · [2 个来源] · 2026-05-01 06:11

研究人员开发了LIMSSR，一个新颖的多模态学习框架，解决了训练过程中缺失数据的问题。与假设数据完整的先前方法不同，LIMSSR利用大型语言模型（LLMs）通过提示引导的插补和融合来推断缺失信息。这种方法旨在通过避免直接重建和减轻幻觉来提高多模态任务中的数据效率。 AI

影响通过利用LLMs在训练期间处理缺失数据，引入了数据高效多模态学习的新范式。

排序理由介绍多模态学习新框架的学术论文。

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CV TIER_1 English(EN) · Huangbiao Xu, Huanqi Wu, Xiao Ke, Yuxin Peng · 2026-05-04 04:00

LIMSSR：LLM驱动的训练时非完整多模态观测下的序列到分数推理

arXiv:2605.00434v1 Announce Type: new Abstract: Real-world multimodal learning is often hindered by missing modalities. While Incomplete Multimodal Learning (IML) has gained traction, existing methods typically rely on the unrealistic assumption of full-modal availability during …
arXiv cs.CV TIER_1 English(EN) · Yuxin Peng · 2026-05-01 06:11

LIMSSR：LLM驱动的训练时非完整多模态观测下的序列到分数推理

Real-world multimodal learning is often hindered by missing modalities. While Incomplete Multimodal Learning (IML) has gained traction, existing methods typically rely on the unrealistic assumption of full-modal availability during training to provide reconstruction supervision o…