发表机构
Korea Electronics Technology Institute (KETI)（韩国电子技术研究院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对多模态情感分析中模态不完整导致性能下降的问题，提出基于语义感知完整性估计的重建方法，并辅以稳定多任务训练策略，在三个基准数据集上实现更精确的情感预测。
AI 中文摘要
近年来，多模态情感分析研究日益采用以文本为中心的方法，以利用文本模态中固有的丰富情感信息。然而，这些方法在现实场景中常因数据部分缺失或噪声而遭受推理性能下降，尤其是当与情感相关的线索缺失时。为解决此问题，我们提出一种新的完整性估计方法，用于量化不完整数据中保留的情感相关信息程度，以指导缺失语义的重建。此外，我们提出一种训练策略，在联合优化情感预测和完整性估计的同时稳定多任务学习。在三个基准数据集上的大量实验和深入分析表明，所提方法能够实现更准确的语义重建，从而带来更精确的情感预测。
英文摘要
Recent multimodal sentiment analysis studies increasingly adopt text-centric fusion approaches to exploit the rich sentiment information inherent in the textual modality. However, these approaches often suffer from performance degradation during inference due to partially missing or noisy data in real-world scenarios, especially when sentiment-related cues are missing. To address this issue, we introduce a new completeness estimation approach that quantifies the degree of sentiment-relevant information preserved in incomplete data to guide the reconstruction of missing semantics. Furthermore, we propose a training strategy that stabilizes multi-task learning while jointly optimizing sentiment prediction and completeness estimation. Extensive experiments and in-depth analyses on three benchmark datasets demonstrate that the proposed approach enables more accurate semantic reconstruction, leading to more precise sentiment prediction.
CommentsAccepted to the Findings of EMNLP 2026
